Roles of the DOCK-D family proteins in a mouse model of neuroinflammation

The DOCK-D (dedicator of cytokinesis D) family proteins are atypical guanine nucleotide exchange factors that regulate Rho GTPase activity. The family consists of Zizimin1 (DOCK9), Zizimin2 (DOCK11), and Zizimin3 (DOCK10). Functions of the DOCK-D family proteins are presently not well-explored, and the role of the DOCK-D family in neuroinflammation is unknown. In this study, we generated three mouse lines in which DOCK9 (DOCK9−/−), DOCK10 (DOCK10−/−), or DOCK11 (DOCK11−/−) had been deleted and examined the phenotypic effects of these gene deletions in MOG35–55 peptide-induced experimental autoimmune encephalomyelitis, an animal model of the neuroinflammatory disorder multiple sclerosis. We found that all the gene knockout lines were healthy and viable. The only phenotype observed under normal conditions was a slightly smaller proportion of B cells in splenocytes in DOCK10−/− mice than in the other mouse lines. We also found that the migration ability of macrophages is impaired in DOCK10−/− and DOCK11−/− mice and that the severity of experimental autoimmune encephalomyelitis was ameliorated only in DOCK10−/− mice. No apparent phenotype was observed for DOCK9−/− mice. Further investigations indicated that lipopolysaccharide stimulation up-regulates DOCK10 expression in microglia and that microglial migration is decreased in DOCK10−/− mice. Up-regulation of C–C motif chemokine ligand 2 (CCL2) expression induced by activation of Toll-like receptor 4 or 9 signaling was reduced in DOCK10−/− astrocytes compared with WT astrocytes. Taken together, our findings suggest that DOCK10 plays a role in innate immunity and neuroinflammation and might represent a potential therapeutic target for managing multiple sclerosis.

The dedicator of cytokinesis (DOCK) proteins are atypical guanine nucleotide exchange factors (GEFs) 2 that regulate Rho GTPase activity. To date, 11 members (DOCK1-11) have been identified, and they are divided into four families: DOCK-A (DOCK1, DOCK2, and DOCK5), DOCK-B (DOCK3 and DOCK4), DOCK-C (DOCK6, DOCK7, and DOCK8), and DOCK-D (Zizimin1 (DOCK9), Zizimin3 (DOCK10), and Ziz-imin2 (DOCK11)) (1-3). The DOCK proteins in the same family share similar sequence homology, and although DOCK-A, DOCK-B and DOCK-C families are relatively well-studied, the information on the function of the DOCK-D family is very limited. The DOCK-D family proteins all activate the Rho GTPase Cdc42, with DOCK10 also activating Rac1 (4), and they all possess N-terminal PH domains, without the SH3 domain or the proline-rich region that are common features for DOCK-A and DOCK-B. DOCK9 mutations are found in patients with keratoconus (5), an eye disease with progressive thinning of the cornea leading to visual impairment, and it contributes to invasive activity of glioblastoma cells (6). It has also been reported that DOCK9 interacts with DOCK4 and promotes remodeling of lateral organization of endothelial cells and lumen morphogenesis (7). It is highly expressed in the brain (8); interestingly, DOCK9 has been identified as a novel interactor of ␣-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid receptors in the brain (9) and regulates neuronal trafficking of these receptors, suggesting its role in synaptic plasticity (10). Overexpression of DOCK9 promotes dendrite growth in the hippocampus (11), indicating that it plays an important role in the central nervous system (CNS).
DOCK10 is expressed in the brain, spleen, and thymus (8), suggesting that it may play a role in neuroinflammation. DOCK10 is up-regulated in B cells by stimulation with interleukin-4, an effect that is not observed with DOCK9 or DOCK11 (12). Studies with DOCK10 Ϫ/Ϫ mice showed that DOCK10 plays a role in B-cell maturation and function (13,14). However, mice with B cell-specific deletion of DOCK10 showed normal development and mild decline in B-cell activation (15), suggesting that the role of DOCK10 in B cells may be small. Recent studies based on transcriptome analyses demonstrated that DOCK10 Ϫ/Ϫ mice live longer than WT mice (16), indicating that DOCK10 is involved in aging processes.
DOCK11 is highly expressed in the spleen and thymus, suggesting its role in the immune system (8). Not very much information is available at present, but deficiency in DOCK11 causes abnormalities in B cells (14), and expression of DOCK11, but not DOCK9, is decreased in aged mice (17,18), suggesting a role of DOCK11 in aging. Taken together, the tissue distribution of DOCK9, DOCK10, and DOCK11 and their functional features, though information is limited, indicated the possibility that these proteins play a role in immune disorders; however, whether any of the DOCK-D family proteins play a role in neuroinflammation or not is unknown.
Multiple sclerosis (MS) is a progressive neuroinflammatory demyelinating disease of the CNS. Experimental autoimmune encephalomyelitis (EAE) is a classical model that shares clinical and pathological characteristics with MS. It is widely used to study pathogenesis of MS and explore therapeutic targets (19,20). There is vast evidence of a role of the adaptive immunity in pathogenesis of MS and EAE, but innate immunity, involving microglia/macrophages and astrocytes, also plays an important role in both triggering and amplifying neuroinflammation (21)(22)(23). The Rho GTPases regulate important functional responses in innate immunity via actin dynamics, and Cdc42 is a key regulator of astrocytic cell polarity and migration following injury (24,25). In this study, we generated three novel mouse lines with deletion in DOCK9, DOCK10, or DOCK11, and examined the role of the DOCK-D family proteins in neuroinflammation using a mouse model of MS.

Deficiency in the DOCK-D family member has no or little effect on the development of neural and immune cells
To elucidate the functions of DOCK-D family members in detail, we generated the DOCK-D family knockout mice, namely DOCK9-deficient (DOCK9 Ϫ/Ϫ ), DOCK10-deficient (DOCK10 Ϫ/Ϫ ), and DOCK11-deficient (DOCK11 Ϫ/Ϫ ) mice ( Fig. 1, A-C). Examples of genotyping for each knockout mouse line, using tail DNA, are also shown ( Fig. 1, D-F). We first investigated the effects of the deficiency in the DOCK-D family member on the development of the CNS and immune system.
To this aim, we first examined the retina, optic nerve and spinal cord to determine whether neural cells and myelin are affected by the lack of the DOCK-D family. Retinal cross-sections of adult WT, DOCK9 Ϫ/Ϫ , DOCK10 Ϫ/Ϫ , and DOCK11 Ϫ/Ϫ mice were stained with an anti-GLAST antibody for Müller glial cells and an anti-NeuN antibody for detection of retinal ganglion cells (top row in Fig. 2A). Both kinds of staining showed no difference among the mouse lines. Moreover, myelin basic protein (MBP) or MBP/NeuN staining of the optic nerves or spinal cord sections, respectively, also revealed no difference among the mice (middle and bottom rows in Fig. 2A). We also examined the expression of astrocytes, microglia, neurons, and myelin in the spinal cord at P10 but found no abnormalities in DOCK9 Ϫ/Ϫ , DOCK10 Ϫ/Ϫ , and DOCK11 Ϫ/Ϫ mice (Fig. S1), indicating that development of neural cells and myelin was not affected by DOCK-D deficiency. We also examined whether there was any deficit in nervous function in these mice. For this, we assessed visual function by multifocal electroretinogram (mfERG) and motor performance by a rotarod test. We found no impairment in visual function or motor performance by the lack of DOCK-D family proteins (Fig. 2, B-E). We next investigated the effects of the deficiency in the DOCK-D family member on the development of the immune system. FACS analysis of the splenocytes revealed no difference in the popu-

The function of DOCK10 in neuroinflammation
lation of CD3 ϩ CD4 ϩ T cells, CD3 ϩ CD8 ϩ T cells, and CD11C ϩ dendritic cells among the mice (Fig. 3, A-C), whereas a mild reduction in the population of B220 ϩ B cells was found in DOCK10 Ϫ/Ϫ mice compared with WT mice (Fig. 3D). Taken together, these results demonstrated that the DOCK-D family has no or little effects on the development of the CNS and immune system.
Because DOCK-D family members are GEFs that are involved in actin cytoskeleton dynamics, we investigated the effects of the deficiency in the DOCK-D family on T-cell proliferation and macrophage migration, both of which require GEF-mediated actin polymerization. The in vitro proliferation assay showed no difference in T cell proliferation with DOCK-D family member deficiency (Fig. 4A). We then investigated the effects of the deficiency in the DOCK-D family member on macrophage migration by using a Boyden chamber assay. We measured the chemotactic response of macrophage migration from the control medium (top chamber) toward medium containing 50 ng/ml of C-C motif chemokine ligand 2 (CCL2; bottom chamber). Interestingly, CCL2-induced cell migration activity was reduced significantly in DOCK10-or DOCK11-deficient macrophages (Fig. 4, B and C), indicating that GEF activity of DOCK10 and DOCK11 was involved in macrophage migration. These results suggest that the DOCK-D family, mainly DOCK10 and DOCK11, may play a role in some diseases in which macrophages are involved.

The function of DOCK10 in neuroinflammation Impaired microglia/astrocyte activation with DOCK10, but not DOCK9 or DOCK11 deficiency
We then investigated the effects of the deficiency in the DOCK-D family member on myelin oligodendrocyte glycoprotein (MOG)-induced EAE, a disease model in which macrophages play an important role in pathogenesis of disease progression. The disease severity of DOCK10 Ϫ/Ϫ EAE mice, as demonstrated by clinical scores, was significantly reduced after day 16 compared with WT EAE mice (Fig. 5A). However, EAE severities of DOCK9 Ϫ/Ϫ and DOCK11 Ϫ/Ϫ mice remained comparable with those of WT mice (Fig. 5A). Pathological analysis of the spinal cords revealed a large amount of infiltrating cells and demyelination in WT, DOCK9 Ϫ/Ϫ , and DOCK11 Ϫ/Ϫ EAE mice, which were reduced in DOCK10 Ϫ/Ϫ EAE mice (Fig. 5, B and D). A similar trend of reduction in infiltrating cells and demyelination was found in the optic nerves of DOCK10 Ϫ/Ϫ EAE mice (Fig. 5, C and E). We further analyzed EAE pathology in WT and DOCK10 Ϫ/Ϫ mice (Fig. 6). We examined the expressions of CD3, B220, iba1, and GFAP in the spinal cord and optic nerve of WT and DOCK10 Ϫ/Ϫ mice at days 0, 12, and 30 after MOG administration. There was a large increase in the number of CD3ϩ cells, B220ϩ cells, iba1ϩ cells, and GFAPϩ cells at day 12 in the spinal cord of both WT and DOCK10 Ϫ/Ϫ mice, with no significant difference between them, but at day 30, the numbers of all these four types of cells were significantly lower in DOCK10 Ϫ/Ϫ mice compared with WT mice (Fig. 6, A and B). Similar trends were observed in the optic nerve (Fig. 6, C and D). In addition, EAE caused weight loss in WT mice by day 30, but this was not observed in DOCK10 Ϫ/Ϫ mice (Fig. S2). We next investigated possible mechanisms contributing to the different effects of the DOCK-D family members on EAE.
It has been reported that expression of DOCK9 and DOCK10, but not DOCK11, were detected in the brain by Northern blotting analysis (8). To find out which cell types in the CNS show high expression of these genes, we assessed their expressions using primary cultured neurons, astrocytes, and microglial cells. Quantitative PCR analysis revealed that DOCK9 mRNA was expressed predominantly in neurons and relatively low expression of DOCK11 mRNA in astrocytes compared with neurons or astrocytes (Fig. 7A). On the other hand, we detected a significantly high expression of DOCK10 mRNA in microglia compared with neurons or astrocytes (Fig. 7A).
DOCK10 is known to be induced by interleukin-4 in chronic lymphocytic leukemia cells and in human peripheral blood B lymphocytes but not T lymphocytes (12,15,26). We investigated whether DOCK-D family gene expressions could be upregulated by an inflammatory stimulus in microglial cells, the primary immune cells of the CNS. The expression of DOCK10 was significantly up-regulated in microglial cells stimulated with lipopolysaccharide (LPS), whereas expressions of DOCK9 or DOCK11 were unchanged (Fig. 7B). Taken together, these results indicate that DOCK10 deficiency might have effects on the function of microglia.
We then investigated the effects of DOCK10 deficiency on microglial migration by using a Boyden chamber assay. We measured the chemotactic response of microglial migration from the control medium (top chamber) toward 50 M ATP (bottom chamber). ATP induced migration of WT microglial cells, which was reduced significantly by DOCK10 deficiency (Fig. 7, C and D), indicating that microglial migration was impaired by DOCK10 deficiency.
We found that DOCK10 mRNA expression in astrocytes was relatively high, unlike DOCK9 or DOCK11 (Fig. 7A). Therefore, we next investigated the effects of DOCK10 deficiency on astrocytic activity by measuring Toll-like receptor (TLR) 4-or TLR9mediated CCL2 production from astrocytes, which has an important role during neuroinflammation. CCL2 production, as measured by ELISA, was up-regulated in WT astrocytes stimulated with LPS or CpG, which was significantly reduced in astrocytes with DOCK10 deficiency (Fig. 7E). Together, these results indicate that activities of not only microglia but also astrocytes were impaired by DOCK10 deficiency, which might partially contribute to the ameliorated severity in DOCK10 Ϫ/Ϫ EAE mice.

Discussion
In this study, we examined the role of the DOCK-D family members in the CNS by generating and studying the knockout mice. To our knowledge, this is the first study to report DOCK9 Ϫ/Ϫ mice. We found that deletion of the DOCK-D family member genes in mice did not cause severe developmental abnormalities, but the proportion of B cells in splenocytes in DOCK10 Ϫ/Ϫ mice was smaller than WT mice, and the migration ability of macrophages is decreased in DOCK10 Ϫ/Ϫ and DOCK11 Ϫ/Ϫ mice. We investigated the roles of the DOCK-D family members in EAE, and the disease severity was reduced in DOCK10 Ϫ/Ϫ mice, but not in DOCK9 Ϫ/Ϫ or DOCK11 Ϫ/Ϫ

The function of DOCK10 in neuroinflammation
mice. We revealed that LPS stimulation induces up-regulation of DOCK10 in microglia, and the migration ability of microglia is decreased in DOCK10 Ϫ/Ϫ mice. In addition, CCL2 up-regulation in astrocytes by inflammatory stimuli was suppressed in DOCK10 Ϫ/Ϫ astrocytes. These findings suggest that, among the DOCK-D family members, only DOCK10 plays a major role in neuroinflammation.
The role of the DOCK GEFs in neuroinflammation has been demonstrated previously. Studies have reported that a lack of DOCK8, which belongs to the DOCK-C family, leads to impaired activity of T cells and microglia, resulting in reduced EAE severity (27,28). In this study, we demonstrated that a lack of DOCK10 does not affect T cells and dendritic cells, but it affects migration ability of macrophages and microglia. Macrophages and microglia play important roles in pathogenesis of EAE (29), and thus this may explain partly for the reduced EAE severity in DOCK10 Ϫ/Ϫ mice. We also showed that LPS-or CpG-induced CCL2 up-regulation in astrocytes is reduced in DOCK10 Ϫ/Ϫ mice. CCL2 produced by astrocytes are required for recruitment of macrophages and promotes EAE disease progression (30). We have previously reported that deletion of ASK1 (apoptosis signal-regulating kinase 1), a mitogen-activated protein kinase kinase kinase, ameliorates EAE disease severity by suppressing CCL2 up-regulation in astrocytes (31). Taken together, DOCK10 may contribute to the pathogenesis of EAE by regulating activities of macrophages, microglia, and astrocytes.
Little is known about the function of DOCK11 at present. One study with DOCK11 Ϫ/Ϫ mice demonstrated that DOCK11 is involved in B-cell development, but not in immune responses (14). In this study, we show that the percentage of B cells in splenocytes is not affected, but the migration ability of macrophages is impaired in DOCK11 Ϫ/Ϫ mice. Expression of microglial DOCK11 is not altered by LPS stimulation, and the apparent phenotype is absent in DOCK11 Ϫ/Ϫ EAE mice. These findings suggest that although the effects may be small, DOCK11 plays a role in the activities of macrophages.
It was surprising to find the lack of phenotype in DOCK9 Ϫ/Ϫ mice. Previous studies reported that DOCK9 is strongly expressed in the developing brain and that DOCK9 promotes dendrite outgrowth in hippocampal neurons (11). It is possible that there is a synaptic impairment that was not detected in our experimental paradigm. Further studies may include behavioral studies to test cognitive functions.
We were also surprised to find a distinctive phenotype only in DOCK10 Ϫ/Ϫ mice, given the similarity in the structures among the DOCK-D family proteins. However, the distributions of these proteins differ in that DOCK9 is mainly expressed in the CNS and DOCK11 is mainly expressed in the immune system, whereas DOCK10 is expressed both in the CNS and immune system (8). In addition, whereas DOCK9 and DOCK11 are Cdc42 GEFs, DOCK10 is a GEF for both Cdc42 and Rac1 (4). Furthermore, we demonstrated that LPS stimulation upregulates DOCK10, but not DOCK9 or DOCK11, in microglia,

The function of DOCK10 in neuroinflammation
suggesting that DOCK10 plays a role in microglia during inflammation. These differences must have caused the reduced severity of EAE only in DOCK10 Ϫ/Ϫ mice, and future studies need to include cell type-specific deletion of DOCK10 in microglia or astrocytes to further elucidate the role of DOCK10 in neuroinflammation.
We demonstrated that TLR ligand-induced production of CCL2 by DOCK10 Ϫ/Ϫ astrocytes was significantly smaller than WT astrocytes (Fig. 7E). However, it is important to note that microglia also produces CCL2 upon inflammatory stimulation; for example, in response to LPS stimulation, microglia may produce 1-5 times more CCL2 than astro-cytes (32,33). For this reason, we checked the purity of our astrocyte culture for microglial contamination using immunohistochemistry. We found that although many GFAP-positive cells are detected, the number of Iba1-positive cells is very small (Fig. S3). Therefore, one may assume that contribution of microglia in our setup is small, although we cannot discard the possibility that the observed effects are partly due to microglia.
In summary, we report that DOCK9 and DOCK11 have little or no effect on the pathogenesis of EAE. However, the disease severity of EAE was ameliorated in DOCK10 Ϫ/Ϫ mice, suggesting differential biological roles of the DOCK-D family proteins

The function of DOCK10 in neuroinflammation
in neuroinflammation and that DOCK10 could be a novel therapeutic target for MS.

Experimental animals
DOCK9 Ϫ/Ϫ , DOCK10 Ϫ/Ϫ , and DOCK11 Ϫ/Ϫ mice were generated by homologous recombination in C57BL/6 mouse ES cells. The animals were treated in accordance with the Tokyo Metropolitan Institute of Medical Science Guidelines for the Care and Use of Animals, and all animal experiments were approved by the Institutional Animal Care and Use Committee of the Tokyo Metropolitan Institute of Medical Science. Genotyping of transgenic mice were determined by PCR using mouse tail genomic DNA and the specific primer sets (Table 1).

Histological analyses
Frozen (10 m thick) or paraffin (7 m thick) tissue sections of the retina, optic nerve or spinal cord were exami- For detection of myelin in the spinal cord and optic nerve of EAE mice, the tissues were embedded in paraffin wax, sectioned at the thickness of 7 m, and stained with Luxol fast blue followed by hematoxylin-eosin. Stained sections were examined using a microscope (BX51; Olympus, Tokyo, Japan) connected to a DP70 camera (Olympus). The images were processed and viewed using a DP manager software (version 2.2.1.195; Olympus).

Multifocal electroretinograms
The mice were anesthetized by intraperitoneal injection of 87.5 mg/kg sodium pentobarbital. The pupils were dilated with 0.5% phenylephrine hydrochloride and 0.5% tropicamide. mfERG were recorded using a VERIS 6.0 system (Electro-Diagnostic Imaging, Redwood City, CA). The visual stimulus consisted of seven hexagonal areas scaled with eccentricity. The stimulus array was displayed on a high-resolution black and white monitor driven at a frame rate of 100 Hz. The secondorder kernel, which is impaired in optic neuritis and optic neuropathy, was analyzed (34 -36).

Rotarod test
Motor performance was examined by the accelerating rotarod test using a rotarod treadmill (Muromachi Kikai Co.,

The function of DOCK10 in neuroinflammation
Tokyo, Japan), according to the manufacturer's instructions. Briefly, the rotating speed was set to accelerate from 5 to 40 rpm over a 300-s period. Each mouse received three trials as training; then after 2-h rest, a mouse was placed on the rotating rod, and the latency to fall and speed at fall were measured.

Chemotaxis assay
Migration assays were performed using Boyden-type 96-well plates (ChemoTx system; Neuroprobe, Gaithersburg, MD). Primary cultured microglia or bone marrow-derived macrophages (BMDMs) from WT, DOCK9 Ϫ/Ϫ , DOCK10 Ϫ/Ϫ , and DOCK11 Ϫ/Ϫ mice were suspended in Dulbecco's modified Eagle's minimal essential medium without serum. The same medium containing CCL2 (10 ng/ml; PEPROTECH, Rocky Hill, NJ) for BMDC and ATP (50 ng/ml; BioVision, Mountain View, CA) for microglia was used to fill each well. Polycarbonate filters with 8-m pores were placed in contact with the liquid, and the cells were dispensed over each well. The cells were seeded on polycarbonate filters and incubated at 37°C (5% CO 2 ) for 3 h. The cells that had remained on the top surface of the filter were wiped with a Kimwipe, and the cells under the filter (migrated cells) were fixed with 3.7% formaldehyde for 10 min. The filter was then stained with a 0.03% crystal violet (Fujifilm, Tokyo, Japan), and individual fields were counted.

Quantitative real-time PCR
Quantitative real-time PCR was performed using an MyiQ single-color real-time PCR detection system (Bio-Rad) with a THUNDERBIRD SYBR qPCR mix (Toyobo, Osaka, Japan). Total RNA for PCR was prepared from tissues from WT mice and from primary cultured cells (neurons, astrocytes, and microglia) (28). Complementary DNA reverse transcribed from total RNA was amplified by using specific primers ( Table 2). The data were normalized to the level of a house-keeping gene, glyceraldehyde 3-phosphate dehydrogenase mRNA.

ELISA for CCL2 measurement
Primary cultured astrocytes were treated with or without LPS (10 ng/ml) or CpG (100 nM) for 16 h. Concentrations of CCL2 in the cell culture media were determined by ELISA (R&D Systems, Minneapolis, MN).

Statistical analyses
The data are presented as mean values, and the error bars represent S.E. The data comparing two groups were analyzed by a two-tailed unpaired Student's t test. Comparison of multiple groups was performed using a Dunnett's test or Tukey-Kramer test. A value of p Ͻ 0.05 was regarded as statistically significant. JMP version 8.0.1 (SAS Institute, Cary, NC) was used for the statistical analyses.