Lipopolysaccharide administration for a mouse model of cerebellar ataxia with neuroinflammation

Most cerebellar ataxias (CAs) are incurable neurological disorders, resulting in a lack of voluntary control by inflamed or damaged cerebellum. Although CA can be either directly or indirectly related to cerebellar inflammation, there is no suitable animal model of CA with neuroinflammation. In this study, we evaluated the utility of an intracerebellar injection of lipopolysaccharide (LPS) to generate an animal model of inflammatory CA. We observed that LPS administration induced the expression of pro-inflammatory molecules following activation of glial cells. In addition, the administration of LPS resulted in apoptotic Purkinje cell death and induced abnormal locomotor activities, such as impaired motor coordination and abnormal hindlimb clasping posture. Our results suggest that intracerebellar LPS administration in experimental animals may be useful for studying the inflammatory component of CA.


Induction of chemoattractants by LPS administration in the mouse cerebellum.
We also examined the production of proinflammatory chemokines, such as MCP-1 and MIP-1α known as chemoattractants for inflammatory cells, in the LPS-injected cerebellum. As shown by western blotting (Fig. 2a), the expression of MCP-1 and MIP-1α in the cerebellum was significantly increased 1 day after LPS injection compared to levels in intact controls, and this increase abated at 7 days. Similar to the western blotting, the double immunostaining showed that the increases in these chemokines were more apparent at 1 day after LPS treatment than 7 days, and that the expression of MCP-1 was mainly observed in microglia, while MIP-1α was mainly observed in astrocytes (Fig. 2b), although some expression of each chemokine was also observed in both astrocytes and microglia in vivo (Fig. 2c), and the mRNA levels of both MCP-1 and MIP-1α (measured by traditional RT-PCR) increased in cultured primary microglia and astrocytes exposed to LPS and LPS + IFN-γ, respectively (Fig. 2d,e).

Effects of LPS on microglial polarization in the mouse cerebellum.
To examine the effects of LPS on the regulation of the microglial phenotype in the mouse cerebellum, we investigated the expression levels of M1-and M2-type microglia 7 days after LPS treatment. In the results of double immunostaining, CD86 (an M1 marker) was highly expressed in Iba1-positive microglia by LPS injection, whereas CD206 (an M2 marker) was rarely expressed in the cerebellum (Fig. 3a). Consistent with the double immunostaining results, western blotting showed that LPS exposure in the cerebellum induced a significant increase in CD86 compared to the level in intact controls with no change in CD206 expression (Fig. 3b). Moreover, in western blotting results, LPS treatment was found to increase the expression of iNOS (an M1 marker) and significantly reduce the expression of IL-10 (an M2 marker) relative to control levels (Fig. 3c).

Motor impairments and Purkinje cell damage following to LPS-induced neuroinflammatory responses.
To investigate the suitability of the LPS-based animal model of inflammatory CA (ICA), we confirmed motor deficits in ICA mice through rotarod testing at weekly intervals for 4 weeks, beginning when the mice were 10-week-old. Retention time on the rotating rod significantly declined, from 572 ± 46 s at 0 week to 351 ± 47 at 4 weeks after LPS injection compared to the performance of intact controls (≥ 580 s) (Fig. 4a). Additionally, we carried out rapid and sensitive phenotype assessments for ataxia-like deficits 4 weeks after LPS injection, as previously described 14 . While there was no mortality in any of the groups (data not shown), LPS-injected mice showed impaired motor coordination (ledge test) and abnormal hindlimb clasping posture compared to intact controls (Fig. 4b). We also evaluated whether ataxia-like motor impairment induced by neuroinflammation was associated with the loss of Purkinje cells, a typical neuropathologic feature of CAs 15 . Relative to control mice, LPS-injected mice exhibited > 50% reduction in calbindin protein, which indicates Purkinje cell death, in the cerebellum 7 days after LPS injection (Fig. 4c). As shown by western blotting and double immunostaining, the activation of caspase-3 showed a time-dependent increase in the LPS-injected cerebellum compared to that of intact controls ( Fig. 4d; Suppl. Fig. S1), and cleaved-caspase-3 was observed in cerebellar Purkinje cells following LPS administration in vivo (Fig. 4e).

Discussion
CAs are neurological conditions resulting from cerebellar inflammation or damage due to various causes 1,3,4 . Recent global studies on ataxia report an overall prevalence rate of 26 cases in every 100,000 children 16 . The causes of CA can be typically divided into three main categories: hereditary, idiopathic, and acquired 17 . A recent accumulation of clinical and preclinical evidence supports the hypothesis that cerebellar inflammation plays a significant role in the progression of CAs. For example, neurotoxic inflammatory responses, involving the production of inflammatory molecules by activated microglia and astrocytes, have been observed in patients and in animal models of hereditary CA, especially spinocerebellar ataxia 6,7 . In addition, some studies suggest that inflammatory responses induced by viral infection or toxic exposures (exposure to alcohol, drugs, or environmental toxins) can cause CA because the cerebellum is especially vulnerable to poisoning and intoxication 8,9,18 . These research findings suggest that the upregulation of inflammatory responses in the cerebellum may be an important mechanism in the progression of CAs and that the control of neuroinflammation may be a potential therapeutic target for the management of CAs.
The development of effective targeted therapies against CA has been hindered by its complex and wide-ranging causes 3 and the lack of suitable animal models. Here, we evaluated the suitability of LPS as a potent promoter of inflammation in the cerebellum to develop a mouse model of ICA. We observed that mice with LPS-induced (for microglia) and GFAP (for astrocytes) in the cerebellum 1 day and 7 days after LPS injection. Iba1: ** p = 0.008 and *** p < 0.001 versus CON, $$$ p < 0.001 versus PBS (7 days), ## p = 0.004 versus PBS (1 day) (n = 3 per group, F(4,10) = 26.219). GFAP: *** p < 0.001 versus CON, $$$ p < 0.001 versus PBS (1 day), ### p < 0.001 versus PBS (7 days) (n = 3 per group, F(4,10) = 150.249) (b) The expression levels of proinflammatory cytokines, such as IL-1β and TNFα, in the cerebellum 1 day and 7 days after LPS injection. IL-1β: ** p = 0.002 versus CON, $ p = 0.014 versus PBS (7 days), ## p = 0.003 versus PBS (1 day) (n = 3 per group, F(4,10) = 11.034). TNFα: ** p = 0.0027 and * p = 0.035 versus CON (n = 3 per group, t(4) = − 6.632 and t(4) = − 3.136). (c) A sagittal histologic section of the mouse cerebellum. The box shows the cerebellar cortex about which the section shown in (d) and (e) was taken. We use "MCL" to represent the molecular cell layer, "PCL" to represent the Purkinje cell layer, and "GCL" to represent the granule cell layer. (d) The expression pattern between microglia (green) and proinflammatory cytokines (IL-1β and TNFα; red) in cerebellum 1 day and 7 days after LPS injection (Scale bar = 50 μm). (e) The expression patterns of astrocytes (green) and proinflammatory cytokines (IL-1β and TNFα; red) in the cerebellum at 1 day and 7 days after LPS injection. IL-1β and TNFα expression was significantly increased by activated microglia in the cerebellum after LPS injection compared to levels in the controls, but were rarely colocalized with astrocytes (arrowheads). Scale bar = 50 μm. All values are expressed as the mean ± standard deviation (SD). Statistical significance tested with one-way ANOVA with Tukey's post-hoc analysis (a and b; IL-1β) and two-sided paired t-test (b; TNFα). The membranes were initially cut according to the molecular weight size of proteins. The images of blots were displayed in cropped format, and original blots of (a) and (b) with cropped demarcation of membranes are shown in Supplementary Fig. 2 www.nature.com/scientificreports/ www.nature.com/scientificreports/ CA develop severe cerebellar inflammation associated with glial activation, resulting in the increased expression of the proinflammatory molecules IL-1β, TNFα, MCP-1, and MIP-1α (Figs. 1, 2). IL-1β and TNFα were mainly produced by activated microglia (Fig. 1) and MCP-1 and MIP-1α, which are known to possess chemoattractant activity and which increased soon after LPS treatment, were predominantly produced by microglia and astrocytes, respectively (Fig. 2). The upregulation of chemoattractants such as MCP-1 and MIP-1α following an inflammatory stimulus plays an important role in the trafficking of microglia toward the site of damage. For instance, increased expression of MCP-1 after hypoxic-ischemic injury facilitates the migration of microglia to injury sites 19 , and the upregulation of MIP-1α induces microglial accumulation in the infarcted brain 20 . Together with this previous evidence, our results suggest that the production of MCP-1 and MIP-1α following LPS treatment might contribute to glial migration and increased inflammatory responses. In neurodegenerative disorders, there are two primary types of activated microglia with different functions: the M1 phenotype producing pro-inflammatory cytokines and the M2 phenotype producing anti-inflammatory mediators such as IL-10, arginase-1, and transforming growth factor-β, which inhibit pro-inflammatory responses, improve neuronal repair, and promote phagocytosis of abnormal proteins 21,22 . The cell surface markers and cytokine expression levels of M1-and M2-type microglia in the LPS-exposed cerebellum indicate that many microglia were activated and polarized toward an M1 phenotype after LPS exposure, demonstrated by the increases in the levels of CD86 and iNOS and by the decreases in the levels of IL-10 (Fig. 3). Consistent with the induction of the proinflammatory molecules, such as IL-1β and TNFα (Fig. 1), these results suggest that LPS administration promote a shift in activated microglia, resulting in the induction of inflammatory effects in the LPS-treated cerebellum.
Accumulating evidence demonstrates that the loss of Purkinje cells, which is a typical pathological feature of CAs, can be associated with inflammation in various types of CA 1,7,23 . Moreover, Purkinje cell loss and dysfunction are the most frequent features in patients and animal models with ataxic symptoms 7,24 . In addition to inflammatory responses of activated glia in the cerebellum, we observed the apoptotic death of Purkinje cells in the LPS-treated cerebellum, and injured mice showed evidence of motor dysfunction, such as impaired coordination, loss of balance, and abnormal hindlimb clasping postures (Fig. 4).
In conclusion, we examined the potential of an intracerebellar injection of LPS to develop a mouse model of ICA. LPS-induced ICA mice displayed impairments of balance and motor control, indicating cerebellar dysfunction, and these features were accompanied by the upregulation of proinflammatory molecules via glial activation, www.nature.com/scientificreports/ followed by apoptotic death of Purkinje cells (Fig. 4f). Thus, these results demonstrate that the LPS-treated mice manifest inflammatory responses and typical symptoms of patients with CA, suggesting their possible utility as an experimental animal model of ICA. The needle was left in place for an additional 5 min to limit reflux along the injection tract. Following injection, mice were moved into a cage with a warming pad to recover from surgery.

Primary glial cultures.
Mixed primary glial cell cultures were prepared from 2-days-old C57BL/6 mice by chopping and mechanically disrupting the whole brain using a nylon mesh, as described previously 26 . Briefly, the cells were grown in DMEM with 10% heat-inactivated fetal bovine serum (FBS, Gibco BRL, Grand Island, NY, USA) and 1% penicillin-streptomycin (Gibco BRL, Grand Island, NY, USA) at 37 °C in a humidified atmosphere of 5% CO 2 . Culture media were replaced with fresh media after the initial 5 days and then every 3 days. Primary microglia were obtained from mixed glial cultures via mild trypsinization. After culture for 3 week, mixed glial cells were treated for 20 min with a trypsin solution (0.25% trypsin-EDTA, Gibco BRL, Paisley, Scotland) diluted 1:3 in serum-free DMEM. This resulted in the detachment of astrocytes, whereas microglia remained attached to the bottom of the culture flask. The detached astrocytes were collected by centrifugation at 5,000 × g for 10 min for primary astrocytes cultures and the remaining primary microglia were used for experiments. The collected microglia and astrocytes were grown and maintained in DMEM with 10% FBS and 1% penicillin-streptomycin.  . All values are expressed as the mean ± standard deviation (SD). Statistical significance tested with two-sided paired t-test (a), Mann-Whitney U statistic (b), and one-way ANOVA with Tukey's post-hoc analysis (c and d). (e) Detection of apoptotic Purkinje cells (arrowheads) in the cerebellum 7 days after LPS injection, using immunofluorescence analysis of cleaved caspase-3. We use "MCL" to represent the molecular cell layer, "PCL" to represent the Purkinje cell layer, and "GCL" to represent the granule cell layer. Insets delimited by the solid line in the panels on the right show high magnification portions of the image. (Scale bar = 100 μm). (f) Schematic representation of ataxia symptoms induced by neuroinflammation in the LPS-exposed cerebellum. The membranes were initially cut according to the molecular weight size of proteins. The images of blots were displayed in cropped format, and original blots of (c) and (d) with cropped demarcation of membranes are shown in Supplementary Fig. 2. www.nature.com/scientificreports/ Behavioral tests. The rotarod testing began 1 day before the LPS injection and it was then conducted weekly for 4 week. For the simple composite phenotype scoring system described below, behavioural testing was performed once, 4 week after LPS injection.

Measurement of MCP-1 and MIP-1α by traditional RT-PCR in glial cultures.
Rotarod test. The rotarod test was used to evaluate motor coordination and balance according to the method described by Zhang et al. 28 . Experimental mice were carefully placed on a rotating rod weekly for 4 week. The rotational speed was linearly increased from 4 to 40 rpm, and then remained at this speed for the remaining 5 min. The total time that the mice could stay on the rotating rod was recorded. To avoid muscle fatigue, each mouse had 10 min of rest between each trial.
Simple composite phenotype scoring system. To quantify disease severity in the LPS-induced mouse model of CA, a previously described scoring system combining a ledge test and hindlimb clasping test was used with modifications 14 . The scoring was performed at 14 week of age. All tests were scored on a scale of 0-3 and combined into a composite phenotype score of 0-6: 0 indicates absence of the relevant phenotype, and 3 indicates the most severe symptoms. Each test was performed in triplicate.
The ledge test Imbalance and incoordination were scored using a ledge test. A score of 0 was applied if a mouse typically walked along the ledge without losing its balance. A score of 1 was applied if a mouse walked in an asymmetrical posture along the ledge. A score of 2 was applied if a mouse lost its footing while walking along the ledge. A score of 3 was applied if a mouse did not effectively use its hind legs.
Hindlimb clasping test Hindlimb clasping has been used as a marker of disease progression in CA mouse models 29 . A score of 0 was applied if the hindlimbs were consistently splayed outward, away from the abdomen. A score of 1 was applied if one hindlimb was retracted toward the abdomen for more than 50% of the time measured. A score of 2 was applied if both hindlimbs were partially retracted toward the abdomen for more than 50% of the time measured. A score of 3 was applied if the hindlimbs were entirely retracted and touching the abdomen for more than 50% of the time measured.