A comprehensive study on long-term injury to nigral dopaminergic neurons following intracerebroventricular injection of lipopolysaccharide in rats

Authors

  • Yan Zhou,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Yanxin Zhang,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Junquan Li,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Fengyue Lv,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Yongmei Zhao,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Deyi Duan,

    1. Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author
  • Qunyuan Xu

    Corresponding author
    • Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing, China
    Search for more papers by this author

Address correspondence and reprint requests to Qunyuan Xu, Beijing Institute for Neuroscience, Beijing Center of Neural Regeneration and Repairing, Key Laboratory for Neurodegenerative Diseases of the Ministry of Education, Capital Medical University, Beijing 100069, China. E-mail: xuqy@ccmu.edu.cn

Abstract

Parkinson's disease (PD) is characterized by selective and progressive degeneration of dopaminergic neurons in the substantia nigra (SN). Lipopolysaccharide (LPS) can induce chronic inflammation and has been widely used to study the pathogenesis of PD. In this study, a single intracerebroventricular injection of LPS was used to induce neurotoxic effects on dopaminergic neurons in Sprague–Dawley rats. The long-term neurotoxic effects of LPS were evaluated at different time points. Microglia were activated in the hippocampus and striatum at 4 weeks, and in the SN at 24 weeks. Astrocytes were activated in the hippocampus and nigrostriatal system at 2 and 24 weeks. The expression of brain-derived neurotrophic factor in the SN increased at 4 weeks and decreased after 12 weeks, and tyrosine hydroxylase-positive neurons in the SN were shown to have an atrophic appearance, with cell loss evident after 24 weeks. Phospho-α-synuclein expression, a reflection of parkinsonian pathogenesis, increased at 12 weeks, and peaked at 24 weeks. Abnormal motor behavior appeared at 16 weeks and lasted up to 48 weeks. These results indicate that microglia are activated for several months after a single, low dose injection of LPS, which eventually results in progressive and selective damage to dopaminergic neurons in the SN.

Abbreviations used
5-HT

serotonin

6-OHDA

6-hydroxydopamine

AChase

acetylcholine esterase

BDNF

brain-derived neurotrophic factor

DAB

3,3′-diaminobenzidine

GFAP

glial fibrillary acidic protein

LPS

lipopolysaccharide

MPTP

1-methyl-1,2,3,6-tetrahydropyridine

PD

Parkinson's disease

PVDF

Polyvinylidene difluoride

SDS

Sodium dodecyl sulfate

SN

substantia nigra

TH

tyrosine hydroxylase

TMB

3,3′,5,5′-Tetramethylbenzidine

Parkinson's disease (PD) is mainly characterized by the selective and progressive degeneration of dopaminergic neurons in the substantia nigra (SN) (Olanow and Tatton 1999), however, its etiology remains unknown. In general, mutations of certain molecules, such as α-synuclein, parkin, and ubiquitin C terminal hydrolase L1 (Mouradian 2002), have been reported to correlate with early on-set and familial PD, which represents only a very small fraction of the incidence of the disease, while the majority of PD cases (sporadic PD) are thought to be linked to environmental toxins (Logroscino 2005), mitochondrial dysfunction (Dauer and Przedborski 2003), and oxidative stress (Przedborski et al. 2003). Among the various factors that may cause PD, inflammation in the brain has garnered much interest because of certain clinical investigations showing an increase in activated microglia and high levels of inflammatory factors (such as tumor necrosis factor-α, interleukin-1β, and nitric oxide) in the nigrostriatal system from post-mortem analyses of PD patients (Mogi et al. 1995). Meanwhile, the on-set of PD, in some cases, has been reported to be associated with head trauma or encephalitis, suggesting an inflammatory component is involved in the disease process (McGeer et al. 2001). In addition, numerous activated microglia and high levels of inflammatory factors were also observed in the brains of animal models of PD, including the 6-hydroxydopamine (6-OHDA) and 1-methyl-1,2,3,6-tetrahydropyridine (MPTP) injury models (Kohutnicka et al. 1998; Li et al. 2005). Moreover, usage of anti-inflammatory drugs has been shown to attenuate toxin-induced PD (Li et al. 2005). These studies all indicate the involvement of inflammation in the neurodegeneration associated with PD.

Lipopolysaccharide (LPS), an endotoxin produced by gram-negative bacteria, is a strong inducer of inflammation. Injection of LPS into the SN (Zhou et al. 2005) or striatum (Hunter et al. 2009) can induce inflammation in the brain and has been used as another tool to produce animal models of PD. These models have a similar phenotype to PD patients, such as bradykinesia and loss of dopaminergic neurons in the SN, however, studies using these models have not been able to explain the role of inflammation in causing PD pathology. This may be because of two factors. One, the LPS injections in situ may have directly damaged the nigrostriatal system. Therefore, loss of dopaminergic neurons in the SN of these models may not solely be because of inflammation, which is inconsistent with the characteristic selective injury of dopaminergic neurons in PD. Second, the observation periods for pathological changes in these models were relatively short, only several months at most (Ling et al. 2006), which does not allow the delayed and progressive worsening of the disease to be monitored, another characteristic of PD.

To address these limitations, we proposed a new method to produce a rat model of PD, where whole brain inflammation was induced by a single intracerebroventricular injection of LPS at a lower dose, with the survival time prolonged up to 1 year. In this study, we show that a single injection of LPS can induce long-term inflammation in the whole brain, but with selective and progressive injury to dopaminergic neurons in the SN, indicating that this model may better mimic the pathology of PD.

Methods

Animal model and tissue preparation

Adult male Sprague–Dawley rats (n = 162) were obtained from the Experimental Animal Center of Capital Medical University, Beijing, China. Animal protocols followed guidelines established by the NIH and were approved by the Animal Care and Use Committee of Capital Medical University. All rats used in the study weighed 250–300 g and were randomly divided into two groups (n = 81 each): the saline injection group (20 μL of saline) and the LPS injection group (25 μg of LPS, diluted in 20 μL of saline). Animals were housed on 12-h light–dark cycles with free access to food and water. All animals were allowed to acclimatize for at least 2 days before operation. Rats were anesthetized by intraperitoneal (i.p.) injection of 6% chloral hydrate (360 mg/kg) and positioned in a stereotaxic apparatus (Benchmark-900; David Kopf, Tujunga, CA, USA). For infusion of LPS (055:B5, Sigma, St Louis, MO, USA) or saline into the right lateral ventricle, a burr hole was made using the following coordinates: 0.8 mm posterior to bregma and 1.3 mm right lateral to the midline; 3.5 mm ventral to the cerebral dura mater, with the incisor bar set 2.4 mm inferior to the interaural line. Saline (20 μL) or LPS (25 μg; diluted in 20 μL of saline) was stereotaxically delivered to the right lateral ventricle at a rate of 1 μL/min using a Hamilton syringe. After each injection, the needle was left in situ for an additional 20 min to avoid reflux along the injection track. The wounds were then sutured. At different time points (2, 4, 6, 8, 12, 16, 20, 24, and 48 weeks) after the operation, all surviving rats were inspected using a motor behavior test. Seventy-two animals from each group were used to determine brain pathology at 4, 12, 24, and 48 weeks.

For brain pathology, rats were anesthetized and then perfused with warm saline followed by 4% (w/v) paraformaldehyde in 0.1 M phosphate buffer (PB, pH 7.4, 4°C) through the aorta. The brain was removed, post-fixed for 2–8 h at 4°C in the same fixative solution, and cryoprotected for 2–3 days in 30% (w/v) sucrose in PB. Serial brain sections were cut at 40 μm thickness on a cryotome (Leica CM 3500, Wetzlar, Germany). Sections were collected from six animals at each time point in each treatment group. For each brain, every sixth section were collected in 0.01 M PBS, and the first and second sections were processed for microglia immunostaining using OX42 and OX6, respectively, the third section was used to stain for different types of neurons using tyrosine hydroxylase (TH) and serotonin (5-HT), and histochemistry for acetylcholine esterase (AChase) was performed. The fourth section was used for glial fibrillary acidic protein (GFAP) immunochemistry, and the fifth section was used for phospho-α-synuclein immunohistochemistry. The remaining 24 animal brains, three for each time point in each treatment group, were cut into 20 μm-thick serial sections, and mounted on to gelatin-coated slides for Fluoro-Jade B staining. In addition, 64 rat brains for brain-derived neurotrophic factor(BDNF)and GFAP staining at 2, 4, 12, and 24 weeks were harvested immediately after cold saline perfusion, frozen in liquid nitrogen and stored at −80°C before analysis (eight at each time point in each treatment group).

Behavior trajectory analysis

Rats were first placed in a 30 × 30 cm box without a cover for 10 min, and their motor behavior was then recorded over 30 min using a Digital Video Camera Recorder (DCR-HC85E, Sony, Tokyo, Japan). The data were analyzed with Rat Behavior Trajectory Analysis Software (EthoVision, V3.1, Noldus, Holland).

Immunohistochemical staining

The primary antibodies in this study included OX42 (mouse, 1 : 1000; BD, Franklin Lakes, NJ, USA), OX6 (mouse, 1 : 1000; BD), TH (mouse, 1 : 20 000, Sigma), 5-HT (rabbit, 1 : 5000; Sigma), GFAP (rabbit, 1 : 700; Sigma), and phospho–α-synuclein (mouse, 1 : 5000; Wako, OSAKA, Japan).

The immunostaining procedure was as follows: sections were first washed three times for 10 min in PBS, incubated at room temperature for 30 min in PBS containing 0.3% (v/v) Triton X-100, and then pre-treated for 5 min with distilled water containing 3% (v/v) hydrogen peroxide. Afterward, sections were incubated in blocking solution [PBS, 0.3% (v/v) Triton X-100 and 5% (v/v) normal goat serum] for 1 h, and then in the primary antibody for 24 h at 4°C. The sections were then incubated for 2 h with biotinylated secondary antibody at 23°C, followed by 2 h in streptavidin-peroxidase. Afterward, sections were incubated in 3,3′-diaminobenzidine (DAB). PBS rinses (3 × 10 min) were performed between each step. All sections were mounted, dehydrated, and coverslipped.

Histochemistry

AChase staining

Sections were first pre-treated in 0.1 M acidum aceticum buffer for 10 min, and then in acetylcholine esterase staining solution overnight at 4°C. Sections were washed with distilled water and reacted with 10% (w/v) ferricyanatum kalium, followed by distilled water rinses (3 × 2 min), and finally mounted, dehydrated, and coverslipped.

Fluoro-Jade B staining

Frozen sections at a thickness of 20 μm were mounted on to 2% (w/v) gelatin-coated slides, and then air dried on a slide warmer at 50°C for at least half an hour. Slides were first immersed in a solution containing 1% (w/v) sodium hydroxide in 80% (v/v) alcohol for 5 min. This was followed by 2 min in 70% (v/v) alcohol and 2 min in distilled water. The slides were then transferred to a solution of 0.06% (w/v) potassium permanganate for 10 min, followed by a rinse in distilled water for 2 min. After 20 min in the staining solution, the slides were rinsed for 1 min in each of the three distilled water washes. Excessive water was removed by briefly draining the slides vertically on a paper towel. The slides were then placed on a slide warmer, set at approximately 50°C, until they were fully dry. The dry slides were cleared by immersion in xylene for at least 1 min before coverslipping with DPX (Sigma).

Western blotting for GFAP

Dissected brains, including the hippocampus, the substantia nigra and the striatum were homogenized in lysis buffer [80 mM Tris-Cl, 2% (w/v) sodium dodecyl sulfate (SDS), 10% (v/v) glycerol, pH 6.8] and centrifuged at 12 500 g/min for 30 min. Protein concentrations of extracts (supernatants) were measured using the BCA kit (Pierce, Rockford, IL, USA). Extracts (10 μg) in SDS loading buffer were denatured for 5 min at 100°C, separated on an SDS-polyacrylamide gel, and then transferred onto a polyvinylidene difluoride membrane (Millipore, Billerica, MA, USA). The membrane was blocked with 5% (w/v) non-fat milk in tris-buffered saline (10 mM Tris, pH 7.5, 150 mM NaCl) containing 0.5% (v/v) Tween-20 (TBST) for 1 h and then incubated with a rabbit antibody to GFAP (1 : 10 000; PTG, CHI, IL, USA) overnight at 4°C. After three washes with TBST, the membrane was incubated with a goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (1 : 5000, Bio-Rad, Hercules, CA, USA) for 1 h at 23°C. The membrane was then washed three times with TBST. After incubating in ECL western blotting reagent (Perkin-Elmer, Waltham, MA, USA), the membranes were exposed to ECL Hyper-film (Amersham, Piscataway, NJ, USA). The amount of protein was determined densitometrically using Quantity One software (Gel Doc 2000 Imagine System; Bio-Rad).

ELISA for BDNF

The procedures for protein extraction and detection of protein concentration were similar to those used in western blotting. Following extraction of protein, 0.1 mL per well of rat BDNF standard solutions were aliquoted into a pre-coated 96-well plate at concentrations of 2000 pg/mL, 1000 pg/mL, 500 pg/mL, 250 pg/mL, 125 pg/mL, 62.5 pg/mL, and 31.2 pg/mL. Sample diluent buffer (0.1 mL) was added to the control well (Zero well). Diluted sample supernatant (0.1 mL) was added to separate empty wells. All samples were measured in duplicate. The plate was sealed and incubated at 37°C for 90 min. The cover was removed, the plate contents were discarded, and the plate was blotted onto paper towel. Biotinylated anti-rat BDNF antibody (0.1 mL) working solution was added to each well, and the plate was incubated at 37°C for 60 min. The plate was washed three times with 0.1 M PBS. The washing buffer was discarded, and 0.1 mL of prepared ABC working solution was added to each well. The plate was incubated at 37°C for 30 min. The plate was washed 5 times with 0.1 M PBS. Prepared 3,3′,5,5′-tetramethylbenzidine (TMB) color developing agent (90 μL) was added to each well and incubated at 37°C for 8 min (shades of blue were seen in the wells). TMB stop solution (0.1 mL) was added to each well (color changed immediately to yellow), and the absorbance was read at 450 nm using a microplate reader (within 30 min of adding the stop solution).

Neuron counting

According to the stereotaxic rat brain atlas (George Paxinos), serial coronal sections of the brain between 0.2 mm and 2.0 mm posterior to bregma were made, and every seventh section was used to perform histochemical staining for AChase to label acholinergic neurons (six sections in all). The numbers of all acholinergic neurons in the basal nucleus of Meynert were counted at x 40 magnification using Stereo Investigator (MBF Bioscience, Williston, VT, USA) described by JF Abisambra and LJ Blair (Abisambra et al. 2010). The mean diameter of all the acholinergic neurons was evaluated using three randomly selected 400x fields from every section of the basal nucleus using Image-Pro Plus 6.0 (Media, Silver Spring, MA, USA). Serial coronal sections of the brain between 4.4 mm and 6.2 mm posterior to the bregma were made to perform immunohistochemistry of TH neurons on every seventh section (six sections in all). Serial coronal sections of the brain between 6.6 mm and 8.2 mm posterior to bregma were made to perform immunohistochemistry of serotonergic neurons in the dorsal raphe nucleus on every seventh section (seven sections in all). Other procedures for counting the numbers of TH and 5-HT neurons were the same as those for acholinergic neurons.

Statistical analysis

Statistical analyses were performed using SPSS 11.5 (IBM, Armonk, NY, USA). Results were typically expressed as the mean ± SD. Means were compared using the two independent-samples t-test and a two-way anova. A value of p < 0.05 was considered statistically significant.

Results

Long-term decrease in motor behavior

All animals survived after intracerebroventricular injections of a single low dose of LPS (25 μg, diluted in 20 μL saline) or saline into the right lateral ventricle at a rate of 1 μL/1 min, and they did not get sick or show any significant changes in body temperature. Approximately 16 weeks later, rats with LPS infusion appeared to show signs of hypokinesia. The behavior trajectories were recorded at 4, 6, 8, 12, 16, 20, 24, and 48 weeks. Movement speed of rats in the LPS group decreased significantly compared with that in the control group at 16, 20, 24, and 48 weeks (decreased 21.1%, 27.9%, 27.0%, and 44.8%, respectively) (Fig. 1).

Figure 1.

Delayed and time-dependent decrease in movement speed following lipopolysaccharide (LPS) infusion. The movement speed of rats in the LPS group significantly decreased by 21.1%, 27.9%, 27.0%, 44.8% at 16, 20, 24, and 48 weeks, respectively, when compared with the saline control (n = 5, *p < 0.05, **p < 0.01).

Activation of glial cells

Immunohistochemical staining for OX42 and OX6 was performed to assess the activation of microglia. At 4 weeks after LPS injection, there were numerous activated microglia in the striatum and hippocampus, as indicated by increased OX 42 immunoreactivity, larger cell bodies, and shorter and thicker processes. However, OX42-positive microglia were less evident in the saline group. Meanwhile, considerable numbers of activated microglia were seen in the SN at 24 weeks after LPS injection. OX6 immunohistochemistry demonstrated that microglia became activated initially in the striatum and hippocampus, and later in the bordering area of the SN (Fig. 2).

Figure 2.

Microphotographs showing the activities of microglia in different areas of the rat brain at 24 weeks after injection. OX42-positive microglia had larger cell bodies, shorter and thicker processes with increased OX42 immunoreactivity in the lipopolysaccharide (LPS) group (d, e, f). OX6-positive microglia were seen to have larger cell bodies, shorter and thicker processes with increased OX6 immunoreactivity in the LPS group (j, k, l). Scale bar = 500 μm for (a–l), 250 μm for insets.

Activation of astrocytes was observed by immunohistochemical staining for GFAP. GFAP-positive astrocytes in the hippocampus and SN were shown to have thicker and more darkly stained processes in the LPS-injected group than in the saline group at 2 weeks. However, there were no such differences between the saline group and the LPS group at 4 weeks and 12 weeks. Interestingly, activation of astrocytes was seen again at 24 weeks (Fig. 3). These results were verified also by western blot studies. Levels of GFAP in the hippocampus and nigrostriatal system were significantly higher in the LPS group than in the saline group at 2 weeks, and were 2.38 fold higher in the hippocampus, and 2.58 fold higher in the nigrostriatal system, but there were no differences in these brain areas between the two groups at 4 and 12 weeks. The levels of GFAP in the LPS group increased again at 24 weeks, and were 1.19 fold higher in the hippocampus and 1.68 fold higher in the nigrostriatal system, compared with the saline group (Fig. 4).

Figure 3.

Microphotographs showing glial fibrillary acidic protein (GFAP) immunostaining in the substantia nigra (SN) at 24 weeks after injection. The GFAP-positive astrocytes in the SN displayed larger cell bodies, shorter and thicker processes with increased GFAP immunoreactivity in the lipopolysaccharide (LPS) group (b) than in the saline group (a). Scale bar = 500 μm (a and b), 250 μm (insets).

Figure 4.

Expression of glial fibrillary acidic protein (GFAP) was detected by western blotting in the hippocampus (a) and nigrostriatal system (b) at different survival times after injection (n = 4, *p < 0.05, **p < 0.01).

Progressive and selective loss of dopaminergic neurons in the SN

Immunohistochemical staining of TH in the SN showed that there were no significant differences between the saline group and the LPS group at 4 and 12 weeks after surgery. At 24 weeks, however, TH immunoreactivity in the LPS group decreased, and a few TH-positive cells seemed to be lost. At 48 weeks, TH immunoreactivity decreased significantly (Fig. 5a and b), and surviving TH neurons in the SN in the LPS-treated group exhibited a shrunken appearance, rough perimeters, and short dendrites, in sharp contrast to the appearance of TH neurons in the saline injected group. The immunohistochemical analysis did not demonstrate any significant loss of TH neurons within the first 12 weeks after LPS injection. However, a significant loss of TH-positive neurons between 24 and 48 weeks after LPS treatment was observed. At 24 weeks, a loss of 33.5% was observed, and this LPS-induced cell loss increased further to 40.1% at 48 weeks (Fig. 5c).

Figure 5.

Delayed and progressive loss of dopaminergic neurons in the substantia nigra (SN) at 48 weeks after injection. Tyrosine hydroxylase (TH)-positive neurons in the lipopolysaccharide (LPS)-treated group exhibited shrunken cell bodies, rougher perimeters, and shorter dendrites (a and b). There were no significant differences at 12 weeks after surgery. Scale bar = 500 μm (a and b), 250 μm (insets). A loss of 33.5% was observed at 24 weeks, and a progressive loss of 40.1% was observed at 48 weeks (c) (n = 7, **p < 0.01, #p < 0.05).

In addition to the changes of in dopaminergic neurons, the effects of LPS injection on cholinergic neurons in the basal nucleus of Meynert and serotonergic neurons in the dorsal raphe nucleus were also investigated in this study. There were no significant difference in staining intensity or cell number between the control group and the LPS group (Fig. 6), indicating that intracerebroventricular injection of LPS did not significantly induce a delayed loss of cholinergic neurons in the basal nucleus of Meynert and serotonergic neurons in the dorsal raphe nucleus.

Figure 6.

Effects of intracerebroventricular lipopolysaccharide (LPS) injection on cholinergic neurons and serotonergic neurons at 48 weeks after injection. Microphotographs show histochemistry of AChE for cholinergic neurons in the basal nucleus of Meynert (a and b) and serotonin (5-HT) for serotonergic neurons in the dorsal raphe nucleus (d and e). There were no differences in the number of cholinergic neurons (c) and serotonergic neurons (f) between the NS and LPS group (n = 7 each). Scale bar = 500 μm (a, b, d, e), 250 μm (insets).

BDNF expression levels

ITo determine the relationship between certain neurotrophic factors and cell death of dopaminergic neurons, the expression levels of BDNF in the brain were analyzed by ELISA at different time points after injection. The levels of BDNF protein in the SN increased by 86.3% in the LPS group at 4 weeks compared with that from the saline group. However, levels decreased by 57.8% and 39% at 12 and 24 weeks, respectively (Fig. 7).

Figure 7.

Changes in brain-derived neurotrophic factor (BDNF) expression in the nigrostriatal system at different time points (n = 4, *p < 0.05, **p < 0.01).

Fluoro-Jade B labeling for degenerating neurons

Fluoro-Jade B was used to detect degenerating neurons at early stages. However, no neurons labeled for Fluor-Jade B at any time point in both the LPS and control groups.

Changes in phospho-α-synuclein in the SN

Immunohistochemistry demonstrated that phospho-α-synuclein-positive cells were distributed throughout the brain, including the cerebral cortex, striatum, hippocampus, substantia nigra, and the dorsal raphe nucleus (data not shown). Our data did not demonstrate any difference in immunoreactive intensity and cell number of positive phospho-α-synuclein neurons in the SN between the saline injection group and the LPS injection group at 4 weeks. However, at 12 weeks, phospho-α-synuclein seemed to aggregate and the number of phospho-α-synuclein-positive cells began to increase, these changes reached a peak at 24 weeks (Fig. 8), followed by a decrease at 48 weeks. However, Lewy body-like inclusions were not observed.

Figure 8.

Changes in phospho-α-synuclein in the substantia nigra (SN) at 24 weeks in the saline (a) and lipopolysaccharide (LPS) (b) groups. Scale bar = 500 μm (a and b), 250 μm (insets).

Discussion

Idiopathic PD is an age-related neurodegenerative disease that usually occurs late in life and progresses over many years. The exact cause for the progressive and selective destruction of the nigrostriatal dopaminergic pathway in PD still remains unknown. One intriguing factor that has been increasingly associated with the pathogenesis of PD is inflammation in the brain.

Microglia, the resident immune cells in the brain, play a role in immune surveillance for the CNS. Under normal conditions, microglia display a ramified morphology and are referred to as ‘resting’ microglia, but upon subtle changes in their micro-environment, or as a consequence of pathological changes, they rapidly transform into an activated state displaying a plastic ameboid morphology (Kreutzberg 1996). Activated microglia release pro-inflammatory molecules, such as IL-1β, tumor necrosis factor a (TNF-α) and nitric oxide (NO), and over-yield of these molecules may cause neuronal death (Minghetti and Levi 1998; Hirsch 2000). Furthermore, the activation of microglia, as a result of the initial neuronal death, may initiate a cascade of events leading to progressive neurodegeneration (Streit 2000). This vicious cycle may be responsible for the progressive loss of dopaminergic neurons in the SN, and may be the main pathophysiological process of PD.

A previous study has shown that dopaminergic neurons significantly decreased 3 days after a direct injection of LPS in the SN (Liu et al. 2000) and 7 days after an intrastriatal injection of LPS (Hunter et al. 2007). These methods, using an inflammatory toxin for in situ injection, exhibited a relative delayed toxic effect on dopaminergic neurons when compared with chemical toxins such as 6-OHDA and MPTP. These approaches belong to an acute injury and are not consistent with the characteristic of PD, which is a delayed and progressive process. In this study, a single low dose of LPS was administered via intracerebroventricular injection, inducing global inflammation in the brain, and allowed the effect of inflammation on different brain areas including the region hosting dopaminergic neurons to be monitored. We found that a significant loss of dopaminergic cells in the SN was not observed until 24 weeks after the injection (33.5% lost) with further reductions seen at 48 weeks (40.1% lost). However, cholinergic neurons in the basal nucleus of Meynert and serotonergic neurons in the dorsal raphe nucleus did not seem to be affected, indicating selective damage because of inflammation, in addition to its progressive toxicity.

A 50% loss of dopaminergic cells at 48 w in our control group is indeed inconsistent with the data reported by others, but the trend of dopaminergic neuron loss with aging is actually consistent with the data reported by them (Sanchez et al. 2008). We suppose the reasons why so many neurons were lost may be as follows: First, the SD rats used in different labs may have dissimilar environments and distinct animal population, which can cause certain divergence of animal reactivity to brain inflammation. In fact, there have been inconsistent data about the number of dopaminergic neurons in the same species of animal. A significant loss of the dopaminergic neurons in the SN has been reported (McCormack et al. 2004),for example, in aged squirrel monkey, but no changes observed by Irwin (Irwin et al. 1994). Second, it is reported that trauma can cause the loss of dopamine neurons in human (Bower et al. 2003). This may indicate that the injections through the lateral ventricle in our study led to a bigger trauma in the rat, than that in those reports (Sanchez et al. 2008), as an operation in rat skull was needed before the injection.

The immunohistochemical staining of TH in the SN showed that there were significant differences between the saline group and the LPS group at 24 and 48 weeks but not at 4 and 12 weeks after surgery, the movement speed of rats in the LPS group decreased significantly compared with that in the control group at 16, 20, 24, and 48 weeks, it seems that there are some mismatch between the TH neuron loss and locomotor activity. Generally motor loss occurs only well after TH neuron loss, although We did not detect the TH neuron at the time points of 16 and 20 weeks, the TH neuron may already have significant differences between the saline group and the LPS group at 16 and 20 weeks. Therefore, from our results, we can conclude that the rodent model of inflammation in the brain created by an intracerebroventricular injection of a low dose of LPS mimics the delayed progression and selective injury observed in the SN in PD.

In our experiments, the number of TH neurons began to reduce at 24 weeks after LPS injection. However, cells labeled by Fluoro-Jade B could not be observed at this time point. Fluoro-Jade B is known to be a high affinity fluorescent marker for the localization of neuronal degeneration during acute or early insult (Schmued and Hopkins 2000). It is not proved; however, those vulnerable neurons in a chronic pathology such as Alzheimer's disease (AD) induced by inflammation can be well labeled by Fluoro-Jade B. For instance, Fluoro-Jade B did not stain neurons in double transgenic model of AD in which a clearly progressive neurodegenerative process has been observed in the CA1 region of the hippocampus (Page et al. 2006). The model we used in this study also mimics a chronic inflammatory course, this may be the reason why the Fluoro-Jade B staining is negative in our study.

Actually the functions of neurotrophic factors related to degeneration of neurons in the circuits of the basal ganglia are quite complicated. It is known that the dopamine neurons in the SN and VTA express BDNF (Stewart et al. 2000), and inhibition of the expression of BDNF can result in death of dopamine neurons partially (Ischiropoulos et al. 1992). Partially, elimination of trkB, the high affinity receptor of BDNF, results in reduction of dopamine neuron number in the old mouse (Pennathur et al. 1999). Several post-mortem studies also show that the level of BDNF in the SN of PD patients decreases significantly (Howells et al. 2000). We observed a long-term change of the BDNF level in a pattern of first increasing and then getting decreasing, after a LPS injection. This indicates that the change of BDNF may reflect a response of the brain to the inflammation, in addition to dopamine neuron injury itself. It seems valuable, therefore, to observe the change of BDNF in a chronic course of disease. The BDNF ELISA in this study showed that the levels of BDNF in the SN increased dramatically at 4 weeks, followed by a significant decrease at 12 and 24 weeks after LPS injection. Noxious stimuli are known to cause astrocyte activation, which can generate the expression of BDNF. The initial increase of BDNF in the SN at 4 weeks may be because of direct stimulation by LPS and a result of astrocyte activation. The decrease of BDNF levels at 12 weeks after treatment with LPS may be because of decompensation of astrocytes, which provide certain trophic factors, in particular BDNF. Some studies have reported that dopaminergic neurons can release BDNF by themselves (Howells et al. 2000), and the loss of neurons in the SN may be responsible for the decrease in BDNF in the brain. In our research, however, the decrease of BDNF in the SN preceded the loss of dopaminergic neurons, suggesting that the reduction of BDNF caused by inflammation in the brain may be involved in dopaminergic neuron injury.

As mentioned above, astrocytes can be activated in response to immunologic challenges and brain injury (Tacconi 1998), and consequently they up-regulate the expression of cell-type-specific proteins such as GFAP (Diedrich et al. 1987). Whether the response to brain inflammation from those long-term activated astrocytes benefits neuronal recovery or not is still controversial; however, it is widely believed that reactive astrocytes at the early stage of brain injury have a beneficial effect on neurons by participating in several biologic processes (Tiveron et al. 1992). Reactive astrogliosis in some aspects does interfere with nerve regeneration (Tiveron et al. 1992; Anderson et al. 2006), and certain toxic mediators produced by activated astrocytes may be involved in the pathogenesis of various neurodegenerative diseases.

Western blotting for GFAP in this study did not show any significant differences between the control and LPS group at early stages of inflammation, which was similar to immunohistochemistry results. Using the same techniques, it could be seen that astrocytes began to be activated from week 24. This activation may be induced by injured dopaminergic neurons, because the number of TH-positive neurons decreased significantly at this time point.

Phosphorylation at Ser-129 of α-synuclein in the brain has been shown to play an important role in pathogenesis of PD. Up to 90% of the α-synuclein is phosphorylated in PD (Anderson et al. 2006), whereas only 4% is phosphorylated in normal conditions. Immunohistochemistry for phospho-α-synuclein in this study demonstrated that both the number of positive cells and immunoreactivity were significantly increased in wide areas of the brain at 12 weeks, and reached a peak at 24 weeks. At this time point, dopaminergic cells in the SN, but not other types of neurons in other areas, were degenerating, suggesting a relationship between phospho-α-synuclein and cell injury in the SN in our animal model. However, further examination is required to confirm whether phospho-α-synuclein-positive cells are dopaminergic neurons or astrocytes, because they may have different effects on the pathogenesis of PD.

In summary, the rat model induced by a single low dose of LPS delivered through the ventricle systematically showed that behavioral symptoms and pathological changes mimicked more closely the manifestation of PD than other LPS administration models. Therefore, we believe this chronic brain inflammation model can be more useful for further research on the pathological mechanism of PD, and may provide certain clues for the treatment of PD.

Acknowledgements

This project was supported by the Major State Basic Research Development Program of China (973 Program) (No. 2006CB500700; No. 2011CB504100), the National Natural Science Foundation of China (No.30430280), and the Beijing Committee of Science and Technology (No. Z0005187040311). The authors have no conflicts of interest to disclose.

Ancillary