Neurons promote encephalitogenic CD4+ lymphocyte infiltration in experimental autoimmune encephalomyelitis

Multiple sclerosis (MS) is an autoimmune disease of the central nervous system characterized by neuroinflammation, leading to demyelination and axonal degeneration. Neuronal excitotoxity mediated by Ca2+/calmodulin-dependent protein kinase IIα (CaMKIIα) results in neuronal damage in experimental autoimmune encephalitis (EAE), an animal model of MS. Here, we define a critical role of excitatory neurons in the pathogenesis of CD4+ lymphocyte accumulation in EAE. We silenced the activity of excitatory neurons in a mouse model of targeted EAE using inhibitory designer receptors exclusively activated by designer drugs (DREADD) under a CaMKIIα promoter. Neuronal silencing mitigated clinical disease scores in EAE, reduced the expression of c-fos, Tnfα, Ccl2, and Ccr2 mRNAs in targeted EAE lesions, and prevented the migration of CD4+ lymphocytes towards neurons. Ccl2 shRNA treatment of targeted EAE suppressed the migration of CD4+ lymphocytes and alleviated the motor deficits of EAE. Our findings indicate that neuronal activation in EAE promotes the migration of CCR2+ CD4+ lymphocytes and that neuronal silencing with an inhibitory DREADD alleviates clinical and molecular markers of disease. Neuronal CCL2 is thought to be involved in promoting lymphocytes migration.

www.nature.com/scientificreports www.nature.com/scientificreports/ CaMKIIα (pCaMKIIα) has been observed in the dorsal horn of the spinal column in EAE by quantitative immunohistochemistry, with both small interfering RNA targeting CaMKIIα and CaMKIIαT286A point mutation reducing EAE clinical scores 16 .
Designer receptors exclusively activated by designer drugs (DREADD) can be used to regulate cell activity in a cell-type-specific fashion in freely moving animals 17,18 . The inhibitory DREADD gated by clozapine-N-oxide (CNO), a pharmacologically inert metabolite of the antipsychotic drug clozapine, silences cell activity by inhibitory G-protein (Gi) signaling 17,18 . By using an osmotic minipump to release CNO continuously, the inhibitory DREADD can exert its designated action for a stipulated duration 19 .
In the present study, we aimed to determine how neuronal silencing would vary the migration of CD4 + lymphocytes and the changes in clinical and molecular disease characteristics neuronal silencing would produce.

Results
Inhibitory DREADD attenuates motor deficits in targeted EAE. To suppress neuronal activity, we injected an hM4Di-encoded adeno-associated virus 9 (AAV9) under a CaMKIIα promoter (CaMKIIα-hM4D [Gi]-mCherry, referred to herein as Gi-DREADD) at spinal level Th 9. A CaMKIIα AAV9 tagged with enhanced green fluorescent protein (EGFP; control-DREADD) was used as a control (Fig. 1a). Two weeks after injection, we detected Gi-and control-DREADD labelled with mCherry and EGFP, respectively, in gray matter at spinal level Th 8 (Fig. 1b). We revealed that DREADD expression was restricted to NeuN + neurons, but not to GFAP + astrocytes, in the dorsal column of the spinal cord (Fig. 1c).
To examine the effect of neuronal silencing with Gi-DREADD administered CNO, we generated a targeted EAE mouse model (Fig. 1d). We assessed EAE scores for 28 days after cytokine and pertussis toxin (PTX) injection. Neuronal silencing by Gi-DREADD attenuated motor deficits when compared with control-DREADD administered CNO (Fig. 1e). The maximum and accumulated EAE scores in Gi-DREADD mice were significantly lower than those in control-DREADD mice (Fig. 1f,g). These results demonstrate that neuronal silencing by Gi-DREADD alleviates motor symptoms in targeted EAE.

Inhibitory DREADD suppresses pathological alterations and inflammatory response.
We studied pathological alterations following Gi-DREADD administration to the dorsal column (Fig. 2a, boxed region) of targeted EAE mice on day 17 (peak of disease). Hematoxylin-eosin (H&E) staining showed that Gi-DREADD significantly reduced the number of inflammatory cells (Fig. 2b,c). When compared with control-DREADD, Gi-DREADD also significantly reduced the number of CD4 + lymphocytes (Fig. 2d,e), and accumulation of CD11b + cells (Supplemental Fig. 1). Moreover, Gi-DREADD reduced the area of demyelination, as assessed with immunostaining for myelin basic protein (MBP), and axonal loss, as assessed by immunostaining for SMI-312, a neurofilament marker (Fig. 2f-i). In situ hybridization showed that Ccl2 mRNA was detected in neurons stained with NeuN (Fig. 2j). Expression of Ccl2 was reduced in spinal cord lesions from Gi-DREADD mice compared with control-DREADD mice (Fig. 2j). Ccl2, Ccr2, c-fos, and Tnfα mRNA levels in the lesions of control-DREADD mice were significantly upregulated compared with those of naïve mice, while Gi-DREADD mitigated this upregulation such that mRNA expression approximated to that of naïve mice (Fig. 2k).
Lastly, we assessed the possibility that neuronal CCL2 in targeted EAE is involved in the migration of CD4 + lymphocytes. We studied a migration assay with Ccl2 knockdown in vitro by administration of Ccl2 shRNA-or control shRNA-carrying-AAV9 (Fig. 4a). We found cortical neurons labelled with MAP2 occupied a major portion of Ccl2 shRNA-expressing cells, and a small portion of Ccl2 shRNA-expressing cells was astrocytes labelled with GFAP (Supplemental Fig. 2c,d,f). Similarly, the major portion of control shRNA was expressed in neurons (Supplemental Fig. 2a,b,e). Ccl2 mRNA levels decreased to 12.8% by Ccl2 shRNA treatment (Fig. 4b). Ccl2 shRNA significantly reduced the migration of CD4 + lymphocytes towards embryonic cortical neurons (Fig. 4c,d). These results suggest that knockdown of Ccl2 in neurons might reduce the migration of CD4 + lymphocytes in vitro. We next studied whether neuronal CCL2 promoted CD4 + lymphocytes migration in vivo. Ccl2 knockdown was performed around EAE lesions by injection of Ccl2 shRNA-or control shRNA-carrying-AAV9 into the spinal cord (Fig. 4e). Ccl2 shRNA reduced the expression of Ccl2 mRNA around the lesions (Fig. 4f). Fluorescent in situ hybridization of Ccl2 with RNAscope probes showed that its expression levels in both neurons and astrocytes, labelled with β-tubulin isotype III (Tubb3) and GFAP, respectively, were downregulated by Ccl2 shRNA treatment ( Fig. 4g-j). Ccl2 shRNA suppressed CD4 + lymphocytes accumulation in situ in the targeted EAE mice (Fig. 4k,l). Knockdown of Ccl2 alleviated EAE severity compared with control shRNA (Fig. 4m). Taken together, neuronal CCL2 is thought to be involved in promoting CD4 + lymphocytes migration in targeted EAE mice. Targeted EAE was induced with MOG 35-55 peptide and local injection of TNFα and IFNγ. Seven days after MOG immunization, DREADD-carrying-AAV9 was injected at spinal level Th 9. EAE scores were assessed from day 0 to 28 after cytokine injection. (e) EAE scores throughout the 28 days (control: n = 8, Gi: n = 9; twoway ANOVA followed by the Sidak test). (f) Maximum and (g) accumulated EAE scores (control: n = 8, Gi: n = 9; Student's t-test). Data are presented as mean ± sem.

Discussion
Neuronal excitotoxity mediated by CaMKIIα causes neuronal damage in EAE 16 . In order to alleviate neuronal excitotoxity in the mouse model of MS, we applied Gi-DREADD combined with CNO administration to targeted EAE mice. We found that neuronal silencing in the targeted EAE mice improved EAE clinical scores by suppressing the migration of CD4 + lymphocytes and alleviated demyelination and axonal degeneration (Fig. 4n).
EAE is the most widely used animal model for MS; however, this model causes disseminated inflammatory lesions in the CNS 20 . This property makes it difficult to investigate the effect of regional intervention on inflammatory lesions of EAE. In the present study, we thus used targeted EAE induced by administration of myelin oligodendrocyte glycoprotein (MOG) and PTX and by stereotactic injection of cytokines to generate focal inflammatory lesions. We detected neuronal activation in the inflammatory lesions by c-fos expression. Activated T cells in the EAE inflammatory lesions stimulated extracellular glutamate release to enhance Ca 2+ influx into neurons 14,21 . Ca 2+ accumulation in neurons induced CaMKIIα activation, thereby promoting phosphorylation of the N-methyl-D-aspartate (NMDA) receptor and consequently causing neuronal excitotoxicity 22,23 . Inhibition of CaMKIIα activity in spinal cord neurons alleviates clinical symptoms in the EAE model 16 , suggesting a pathological role of excitatory neurons in EAE. www.nature.com/scientificreports www.nature.com/scientificreports/ In the present study, neuronal silencing induced by inhibitory DREADD mitigated hind-limb paralysis, restored pathological alterations, and attenuated upregulation of Ccl2, Ccr2, c-fos, and Tnfα mRNAs in the spinal cord lesions of targeted EAE. The CCL2-CCR2 pair is critical for expression of EAE 3,5,6 . CCL2 is also produced in embryonic cortical neurons of mice and functions in neuron signaling and development 24 . CCL2 has been shown to attract CCR2-expressing encephalitogenic T cells in migration assays 6,25 . Therefore, we used embryonic cortical neurons to study the activity of neuron-derived CCL2 in a migration assay. During in vitro experiments, embryonic cortical neurons attracted CCR2-expressing CD4 + lymphocytes from EAE mice. Both administration of Gi-DREADD-carrying-AAV9 and Ccl2 shRNA treatment to embryonic cortical neurons suppressed CD4 + lymphocytes migration. Downregulation of neuronal CCL2 would suppress CD4 + lymphocytes migration. Several studies revealed that CCL2 attracted CCR2-expressing encephalitogenic T cells in migration assays 6,25 , and neuronal CCL2 promoted migration of immune cells 7,8 . Activated neurons would attract inflammatory lymphocytes via CCL2 upregulation, and contribute to EAE pathogenesis. In the present fluorescent in situ hybridization study, Ccl2 shRNA treatment reduced the expression of Ccl2 in both neurons and astrocytes in the spinal cord of targeted EAE mice. A previous study clarified the involvement of astrocytic CCL2 to pathophysiology of EAE 13 . We thus assume that astrocytic CCL2 also potently promotes CD4 + lymphocytes migration in target EAE mice.
Previous studies demonstrated that systemic administration of TNFα and local delivery of TNFα by retrovirus-transduced T lymphocytes exacerbated EAE 26,27 . By contrast, anti-TNFα antibody reduced the severity of EAE 28 . TNFα administration upregulated CCL2 expression in cortical neurons 29 . In the current study, activated CD4 + lymphocytes exhibiting high migration activity enhanced TNFα expression. TNFα is also thought to be involved in disease process on EAE.
According to the results of H&E staining and CD4 immunostaining in the current study, CD4 + lymphocytes were a part of cells accumulated in targeted EAE lesion. Targeted EAE lesions show accumulation of T cells, macrophages, microglia, and astrocytes 20 . We found that accumulation of CD11b + cells was reduced after Gi-DREADD in the spinal cord of targeted EAE mice. We assume that neuronal silencing could also reduce accumulation of these immune cells. EAE is initiated when activated autoreactive T cells cross the blood-brain barrier Tubb (magenta), and staining with GFAP antibody (brown) and DAPI (blue) in the spinal cord of (g) control shRNA-or (h) Ccl2 shRNA-injected mice. Scale bar; 100 μm. Quantification of (i) Ccl2-expressing neurons and (j) Ccl2-expressing astrocytes (n = 4, Student's t-test). (k) Representative images of CD4 + lymphocytes in the dorsal column of naïve and control-or Ccl2 shRNA-injected mice and (l) quantification of number of CD4 + lymphocytes (n = 3-5, one-way ANOVA followed by Tukey comparison test). Scale bar; 100 μm. (m) EAE scores (n = 4, two-way ANOVA followed by Sidak test). Data are presented as mean ± sem. (n) Hypothetical illustration. Excitation of CaMKIIα-expressing neurons of targeted EAE mice promotes CD4 + lymphocytes migration which herald demyelination, axonal loss, and consequently motor dysfunction (blue arrows). Neuronal CCL2 upregulated by neuronal excitation would be involved in lymphocytes migration. Neuronal silencing of CaMKIIα-expressing neurons suppresses CD4 + lymphocytes migration via reduction of neuronal CCL2 production and alleviates EAE pathology (black bar and arrows). CST, corticospinal tract.
www.nature.com/scientificreports www.nature.com/scientificreports/ to reach the CNS 6,9 . Deletion of CD4 + lymphocytes by anti-CD4 antibody administration completely protected EAE mice from development of the disease 30. We recognize that neuronal silencing by Gi-DREADD targets on attenuation of lymphocytes infiltration into the CNS at the early stage of the disease.
To our knowledge, we for the first time adopted inhibitory DREADD and a migration assay using embryonic cerebral cortex neurons with CD4 + T cells. This methodology will pave a new way to investigate the pathophysiological role of excitatory neurons in EAE.

Materials and methods
Animals. Eight-week-old male C57BL/6 J mice (Japan SLC) were used in this study. All experimental procedures were approved by the Institutional Animal Care Committee of Osaka University and complied with the guidelines for the care and use of laboratory animals at Osaka University.

Induction of targeted EAE with DREADD.
Targeted EAE was generated according to a previous study 1 .
Seven days after MOG 35-55 immunization, we performed dorsal laminectomy under anesthesia and injected Gior control-DREADD-carrying rAAV9 (1.5 μl, 1.5 × 10 11 gc) at spinal level Th 9. Twenty-one days after injection, we administered a cytokine mixture containing TNFα (750 ng, R&D Systems) and IFNγ (2,500 U, Peprotech) at spinal level Th 8. On the same day, we implanted an osmotic pump containing clozapine N-oxide (CNO, 4 mg/ ml in PBS, Enzo Life Sciences). CNO was infused for 14 days at a flow rate of 2 μg/h. We intravenously administered pertussis toxin (200 ng, List Biological Laboratories) immediately prior to and 48 h after cytokine injection. We assessed daily EAE clinical scores according to the following criteria 1 : 0, no clinical disease; 0.5, partial tail weakness or slight loss of muscle tone; 1, tail weakness; 1.5, loss of tail reflex with hind leg inhibition without gait abnormality; 2, partial hind limb paralysis with gait abnormality; 3, complete hind limb paralysis; 4, front and hind limb paralysis; and 5, moribund state.
The images of H&E-stained sections were acquired with bright field microscopy (Olympus, IX83), while those of immunohistochemical studies were acquired with confocal microscopy (Olympus, FV3000). All images were analyzed using ImageJ software (National Institutes of Health).
In situ hybridization. The animals were perfused with diethylpyrocarbonate (DEPC)-treated ice-cold PBS, followed by DEPC-treated 4% PFA under anesthesia. The tissues were fixed with 4% PFA for 3 days at 4 °C and subsequently transferred to DEPC-treated 30% sucrose in DEPC-treated PBS overnight at 4 °C. Tissues were cut into 25-μm-thick sections using a cryostat. The primer sequence for the Ccl2 RNA probe is shown in Table S2. We integrated PCR product into pCR TM -Blunt II-TOPO ® (Invitrogen). DNA templates were digested with BamH1 or Xho1 for 3 h at 37 °C. Linearized DNA was labelled with DIG RNA labelling mix (Roche) and transcribed by T7 or SP6 RNA polymerase (Roche) for 2 h at 37 °C, followed by lithium chloride precipitation.
Sections were treated with proteinase K for 10 min at RT and acetylated with 0.1 M triethanolamine and 0.25% acetic anhydride for 10 min at RT. After immersion with hybridization buffer for 1 h at RT, the sections were Isolation of CD4 + lymphocytes from the spleen. Targeted EAE mice were anesthetized and perfused with ice-cold PBS. The spleen was dissected and underwent hemolysis. CD4 + lymphocytes were collected using CD4 + T cell isolation kit (Myltenyi) according to the manufacturer's protocol. Subsequently, 5 × 10 6 /ml cells were suspended in RPMI-1640 (Gibco) containing 2% FBS and 1% penicillin-streptomycin. Fig. 3b. To silence neuronal activity, CNO (1 μM in medium) was added to cultured neurons for 30 min at 37 °C. Then, 5 × 10 5 CD4 + lymphocytes isolated from the spleen in 100 μl of medium were transferred to a 5 μm pore size cultured insert (Corning) precoated with 50 μg/ml of fibronectin. Cortical neurons and CD4 + lymphocytes were incubated for 4 h at 37 °C, and then CD4 + lymphocytes on the lower filter surface were fixed with 4% PFA for 30 min at RT. Fixed cells were treated with 0.1% Triton-X and stained with DAPI. Four independent fields were observed with fluorescence microscopy (Olympus, BX5).

Quantitative polymerase chain reaction (qPCR).
We analyzed gene expression of spinal cord tissues in targeted EAE mice and in cultured cortical neurons collected after the migration assay. The procedures were performed as described in a previous study 40 . The primer sequences are described in Table S2 [41][42][43][44] . The relative expression was normalized to 18S rRNA.
shRNA experiment. A pair of 60-nucleotide oligonucleotides encoding the 19-nucleotide Ccl2 shRNA (target sequence: 5′-GAAGTTGACCCGTAAATCT-3′) 8 and turbo red fluorescent protein were designed and cloned into BamH1 and EcoR1 sites of the pAAV-MCS plasmid (Stratagene). Ccl2 shRNA-expressing AAV9 was generated according to the method prepared for DREADD. Throughout the in vivo experiment, the methods for generating targeted EAE and measurement of clinical scores were the same as those used in the DREADD experiment. Ccl2 or control shRNA-carrying-AAV9 was injected at the Th 9 spinal level 14 days after induction with MOG  . For the in vitro experiment, Ccl2 shRNA was administered to cortical neurons isolated from E18 mice 5 days before performing the migration assay. The number of migrating CD4 + lymphocytes was assessed with confocal microscopy (Olympus, FV3000). Successful downregulation of Ccl2 expression was examined by RT-PCR in shRNA-treated neurons. fluorescent in situ hybridization. We used RNAscope probes (Ccl2 311791; Tubb3 423391-C4) designed by Advanced Cell Diagnostics (ACD), and stained four 16-μm cryosections from the spinal cord of mice with ACD RNAscope kits according to manufacturer's instructions.

Statistics.
Results are expressed as mean ± sem. Significance was measured with unpaired t tests, one-way ANOVA followed by the Tukey test, or two-way ANOVA followed by the Sidak or Bonferroni tests. A P value of <0.05 was considered statistically significant. Statistical analyses were conducted using GraphPad Prism 7 (GraphPad software).