Partial suppression of M1 microglia by Janus kinase 2 inhibitor does not protect against neurodegeneration in animal models of amyotrophic lateral sclerosis

Background Accumulating evidence has shown that the inflammatory process participates in the pathogenesis of amyotrophic lateral sclerosis (ALS), suggesting a therapeutic potential of anti-inflammatory agents. Janus kinase 2 (JAK2), one of the key molecules in inflammation, transduces signals downstream of various inflammatory cytokines, and some Janus kinase inhibitors have already been clinically applied to the treatment of inflammatory diseases. However, the efficacy of JAK2 inhibitors in treatment of ALS remains to be demonstrated. In this study, we examined the role of JAK2 in ALS by administering a selective JAK2 inhibitor, R723, to an animal model of ALS (mSOD1G93A mice). Findings Orally administered R723 had sufficient access to spinal cord tissue of mSOD1G93A mice and significantly reduced the number of Ly6c positive blood monocytes, as well as the expression levels of IFN-γ and nitric oxide synthase 2, inducible (iNOS) in the spinal cord tissue. R723 treatment did not alter the expression levels of Il-1β, Il-6, TNF, and NADPH oxidase 2 (NOX2), and suppressed the expression of Retnla, which is one of the markers of neuroprotective M2 microglia. As a result, R723 did not alter disease progression or survival of mSOD1G93A mice. Conclusions JAK2 inhibitor was not effective against ALS symptoms in mSOD1G93A mice, irrespective of suppression in several inflammatory molecules. Simultaneous suppression of anti-inflammatory microglia with a failure to inhibit critical other inflammatory molecules might explain this result. Electronic supplementary material The online version of this article (doi:10.1186/s12974-014-0179-2) contains supplementary material, which is available to authorized users.


Introduction
Amyotrophic lateral sclerosis (ALS) is a devastating disease characterized by progressive degeneration of motor neurons in the brain and spinal cord, resulting in muscle weakness. Although the precise mechanism of ALS remains unknown, inflammatory microglial activation plays an important role in pathogenesis [1][2][3]. The inflammatory molecule IFN-γ, which is primarily produced by Th1 lymphocytes and is a potent activating factor for inflammatory M1 microglia, contributes to the loss of motor neurons in ALS [4,5]. Furthermore, a recent report showed that Ly6c-high inflammatory monocytes are recruited to the spinal cord in mSOD1 G93A mice, and that treatment with anti-Ly6c monoclonal antibodies reduces monocyte recruitment to the spinal cord and ameliorates neurodegeneration in these animals [6].
Janus kinases (JAKs) are centrally implicated in cytokine receptor-mediated cell signaling pathways, which drive a range of myeloid malignancies [7] as well as inflammatory diseases [8]. The JAK family member JAK2 is responsible for transducing signals for several proinflammatory cytokines including IFN-γ and Il-12, as well as for differentiation of myeloid cells [9]. In an animal model of rheumatoid arthritis (RA) and experimental autoimmune encephalitis (EAE), suppression of the JAK pathway ameliorated disease severity by suppressing Th1 cells and deactivating monocytes [10,11]. A growing number of JAK inhibitors have been developed and clinically applied to the treatment of various inflammatory disease including RA, psoriasis, and inflammatory bowel disease [12][13][14]. Although activators of JAK2 such as IFN-γ, Il-6, and Il-12 are reported to be implicated in ALS pathogenesis [3], the role of JAK2 in the ALS-related neuroinflammation remains totally unknown.
Based on these findings, we hypothesized that JAK2 inhibition could ameliorate neurodegeneration in ALS model mice by inhibiting harmful inflammatory processes in microglia/macrophages. To test this idea, we treated transgenic mice overexpressing the familial ALS-associated G93A SOD1 mutation (mSOD1 G93A mice) with R723, an orally active inhibitor of JAK2 [15].

Ethics statements
All animal experiments were conducted in accordance with the guidelines of Osaka University, which specifically approved this study (Permit number: Biken-AP-H21-28-0).
Animals and R723 administration mSOD1 G93A mice were obtained from The Jackson Laboratory and backcrossed with C57BL/6 mice for at least 10 generations. R723 was administered by oral gavage starting on day 90. For the analysis of motor function by rotarod test, weight measurement, and survival, R723 dosing continued until day 120 (70 mg/kg twice daily; 5 days on, 2 days off). To evaluate in vivo pharmacokinetics, plasma and spinal cord tissues were collected at 0.5, 1, 2, and 4 hours post-dose, and R723 levels in plasma and spinal cord tissue were determined by LC/MS/MS.

Nissl staining
Spinal cord sections of mSOD1 G93A mice were prepared as previously described [1]. Every fifth section was collected and stained with cresyl violet.

Statistics
Data are expressed as means ± SEM. Differences in animal weight measurements and rotarod tests were assessed using analysis of variance (ANOVA). Statistical significance in survival experiments was determined using Kaplan-Meier survival statistics. Statistical significance in all other experiments was assessed using the Mann-Whitney U-test. P < 0.05 was considered statistically significant.

Results
To confirm whether expression of inflammatory cytokines was upregulated in the spinal cords of late-stage mSOD1 G93A mice, we evaluated spinal cord mRNA expression of several genes encoding inflammatory molecules. Consistent with a previous report [16], RT-qPCR analysis revealed that the expression levels of IFN-γ, Il-6, Il-12a, and granulocyte macrophage colony-stimulating factor (GM-CSF) increased along with disease progression ( Figure 1A and Additional file 1: Supplementary information). In addition, microglia in the spinal cords of late stage mSOD1 G93A mice (130 days old) had enhanced phosphorylation of JAK2 compared with pre-onset stage mSOD1 G93A mice (70 days old), providing a therapeutic rationale for JAK2 inhibition against ALS ( Figure 1B, C).
To investigate the role of JAK2 pathway in ALS, we used R723, which is a selective small-molecule JAK2 inhibitor originally developed by Rigel Pharmaceuticals Inc, (San Francisco, CA, USA) for the treatment of myeloproliferative neoplasms such as polycythemia vera, essential thrombocythemia and primary myelofibrosis (Additional file 2: Figure S1A) [15]. First, to investigate the drug distribution, we administered R723 by oral gavage to mSOD1 G93A mice and measured concentrations of R723 in serum and spinal cord tissue. R723 had sufficient access to spinal cord tissue (Figure 2A, B)   To further confirm the anti-inflammatory effect of R723, we evaluated the microgliosis and astrocytosis in spinal cord tissue of R723-treated mSOD1 G93A mice. Lectin staining revealed that R723 treatment had suppressed microgliosis in the spinal cords of mSOD1 G93A mice, although it did not affect astrocytosis ( Figure 3A and Additional file 3: Figure S2A). In addition, we evaluated the mRNA expression of inflammation-related and M1/M2 microglia-related genes in spinal cord tissue of R723-treated mSOD1 G93A mice. Consistent with the anti-inflammatory effects of JAK2 inhibitor as previously reported [17], R723 treatment suppressed IFN-γ and iNOS expression dose-dependently, suggesting that the drug exerted anti-inflammatory effects in the spinal cords of mSOD1 G93A mice ( Figure 3B, C and Additional file 4: Figure S3A). In addition, the effect of R723 against iNOS expression was confirmed by immunohistochemical analysis ( Figure 3D). However, R723 had no obvious effects on other inflammatory molecules such as TNF, Il-12b, Il-6, Il-1β, and NOX2. Additionally, there was no significant difference between two groups in the spinal cord expression levels of monocyte chemotactic protein 1 (MCP1) and Ly6c, which are important for the migration and activation of inflammatory monocytes, as well as those of Il-4, arginase, liver (Arg1), chitinase-3-like 3 (Ym1), erythropoietin (EPO), and colony-stimulating factor 3 (CSF3), which are involved in the activation of M2 microglia ( Figure 3E). Unexpectedly, R723 suppressed expression of resistin-like alpha (Retnla), a marker of anti-inflammatory M2 microglia, in spinal cord tissue of mSOD1 G93A mice after 30 days of treatment, although this effect was not evident after 5 days of treatment ( Figure 3F and Additional file 4: Figure S3B). Collectively, these results suggest that oral administration of R723 decreased the number of Ly6c-positive monocytes in peripheral blood and reduced the expression of several inflammatory genes in the spinal cords of mSOD1 G93A mice, leading to suppressed microglial activation. Because R723 suppresses several pathways that seem to be harmful in ALS, we tested whether R723 could ameliorate neurodegeneration in mSOD1 G93A mice. Oral administration of R723 (70 mg/kg, twice daily; 5 days on, 2 days off) to mSOD1 G93A mice was started at 90 days of age and continued until 120 days of age. Motor performance was evaluated by rotarod test, and muscle atrophy was monitored by body weight reduction. Decline in motor performance of the R723-treated mSOD1 G93A mice was compared with that of the vehicle-treated littermates. Throughout the disease process, there was no significant change in rotarod performance or body weight between the two groups ( Figure 4A, B) (P > 0.05 for each time point, ANOVA). Additionally, survival times for R723-treated and vehicle-treated mSOD1 G93A mice were comparable ( Figure 4C) (average survival time; R723 treated group: 155.6 ± 1.8 days (n = 25); vehicletreated group: 155.1 ± 2.2 days (n = 28), P = 0.96, log-rank test). Consistent with these observations, Nissl staining revealed that R723 treatment had led to unaltered motor neuron survival in the spinal cords of mSOD1 G93A mice in both groups ( Figure 4D). Collectively, these results showed that R723 penetrated the spinal cord of mSOD1 G93A mice and suppressed inflammation, but did not affect neurodegeneration in vivo.

Discussion
In this study, we tried to suppress harmful inflammatory processes in mSOD1 G93A mice by treating them with a JAK2 inhibitor. Although R723 was effective in depleting Ly6c-positive monocytes and suppressing IFN-γ and iNOS expression in the spinal cord, the drug did not affect disease progression or survival of these mice.
We examined the expression levels of inflammationrelated genes including TNF, MCP1, Il-1β, and NOX2, which play critical roles in the pathogenesis of ALS [3], and found that they were not reduced after JAK2 inhibition. It is possible that inflammation driven by these molecules masked the effects of reductions in IFN-γ and iNOS expression.
One explanation for the lack of a neuroprotective effect of R723 could be the suppression of Retnla, an M2 microglia-related gene. M2 microglial activation, which is driven by Il-4 and Il-13 produced by Th2 lymphocytes, exerts protective roles in ALS [3]. Recently, another group reported that JAK2 is activated after the recruitment of Il-13 to its receptor, and revealed that Il-13 utilizes the JAK2 signaling pathway [18]. Therefore, we speculate that suppression of Retnla counteracts the anti-inflammatory effects of R723, preventing it from exerting a neuroprotective effect in vivo.
Alternatively, R723 might have inhibited a neuroprotective effect of JAK2. There is a report that suggests JAK2 signaling is implicated in the prevention of neuronal apoptosis in traumatic brain injury [19].
In conclusion, R723 alone was not sufficient to protect against neurodegeneration in mSOD1 G93A mice, although it suppressed the expression of several proinflammatory molecules and depleted monocytes. Based on our results, it is possible that in order to ameliorate neurodegeneration in ALS, we need not only to suppress JAK2 mediated inflammation but also prevent other inflammatory pathways. Furthermore, we may need to activate neuroprotective M2 microglia to alleviate neurodegeneration in ALS. Peripheral blood cells were collected from mSOD1 G93A mice (70 days old mice and 130 days old ones). The following antibodies were used: APC-Cy7-labeled anti-CD11b and FITC-labeled anti-Ly6c. Flow cytometry was performed using a FACS Canto™ II with the Diva ™ software and acquired data were analyzed using the FlowJo software.

Additional files
Additional file 2: Figure S1. R723 is a selective small-molecule JAK2 inhibitor. (A) Chemical structure of R723 is shown.
Additional file 3: Figure S2. R723 had no effect on astrocytosis in the spinal cords of mSOD1 G93A mice. (A) The number of GFAP-positive astrocytes in the spinal cord did not differ between R723-treated mSOD1 G93A mice and vehicle-treated controls. Lumbar sections of the spinal cord were stained with Alexa Fluor 488®-conjugated anti-GFAP antibody. Scale bar =100 μm. Data are representative of three animals.
Additional file 4: Figure S3. R723 had a dose-dependent effect on the suppression of inflammation-related genes. (A) Quantitative RT-PCR analyses revealed that lower dose of R723 (17.5mg/kg, twice a day, 5 days on/2 days off regimen) did not change the expression profiles of IFN-γ, iNOS and Retnla in the spinal cords of mSOD1 G93A mice (n = 3 in lower dose group and n = 4 in other groups). (B) Quantitative RT-PCR analyses in spinal cords of R723-treated mSOD1 G93A mice and vehicle-treated controls were performed after 5 days of treatment (n = 3 in each group). The expression level of iNOS was significantly reduced in the R723-treated group (P = 0.0495). Data are expressed as means ± SEM. *P < 0.05, Mann-Whitney U-test.

Competing interests
YH is an employee of Rigel Pharmaceuticals, Inc. The remaining authors declare that they have no competing interests.