Detection of disease-associated microglia among various microglia phenotypes induced by West Nile virus infection in mice

West Nile virus (WNV) has emerged as a significant cause of viral encephalitis in humans and horses. However, the pathogenesis of the West Nile encephalitis remains unclear. Microglia are activated by WNV infection, and the pathogenic involvement of their phenotypes is controversial. In this study, we examined the diversity of microglia phenotypes caused by WNV infection by assessing various microglia markers and identified disease-associated microglia in WNV-infected mouse brain tissue. Cells positive for general microglia markers such as Iba1, P2RY12, or TMEM119 were detected in the control and WNV-infected brain tissue. The morphology of the positive cells in brain tissue infected by WNV was different from that of control brain tissue, indicating that WNV infection induced activation of microglia. The activated microglia were classified into various phenotypes by investigation of specific marker expression. Among the activated microglia, disease-associated microglia that were positive for CD11c and weakly positive for TMEM119 were detected close to the WNV-infected cells. These results indicate that WNV infection induces activation of diverse microglia phenotypes and that disease-associated microglia may be associated with the pathogenicity of WNV infection in the mouse brain.


Introduction
West Nile virus (WNV) is a member of the family Flaviviridae, genus Orthoflavivirus, species Orthoflavivirus nilense (Schoch et al. 2020), and belongs to the Japanese encephalitis serocomplex (Petersen et al. 2013). In nature, WNV is maintained between mosquitos and birds as vectors and amplifying hosts, respectively. Humans and horses are incidental hosts and are infected with WNV by mosquito bites. The clinical symptoms of WNV infection in humans include febrile illness, fetal meningoencephalitis, and acute flaccid paralysis (Davis et al. 2006). No WNV-specific therapy is available, and vaccines against WNV are only approved for use in horses (Ulbert 2019). This indicates the need to develop an effective and safe treatment and vaccine based on an understanding of the pathogenesis of WNV.
WNV infects neuronal cells in the central nervous system (CNS) and induces neuronal cell death through caspase 3-dependent apoptosis Peng and Wang 2019;Stonedahl et al. 2020). In addition, the capsid protein of WNV inhibits autophagy and induces protein aggregation in WNV-infected neuronal cells, leading to neuronal cell death and inflammation involving activated microglia (Kobayashi et al. 2020;Quick et al. 2014;Stonedahl et al. 2020).
Microglia are the resident immune cells of the CNS. The roles of microglia are involved in homeostasis and host defense against pathogens (Hickman et al. 2018). Under physiological conditions, microglia are morphologically characterized by small and circular soma with an intensive branching process; this is termed the ramified phenotype (Jurga et al. 2020). After recognition of pathogens, microglia become activated; their morphology transforms from the ramified to the amoeboid phenotype, which is characterized by a rounder cell body with short and thick pseudopodia (Chen et al. 2019;Jurga et al. 2020;Quick et al. 2014). Activated microglia are observed in several neurodegenerative diseases, including Alzheimer's disease (AD) and amyotrophic lateral sclerosis (Kenkhuis et al. 2022;Takahashi 2023). In mouse studies, activated microglia degraded Aβ plaque, which is the distinctive pathological hallmark of AD, and inhibited neuronal cell death in the early stages (Feng et al. 2020;Kadowaki et al. 2005). In another study, however, activated microglia chronically produced excess proinflammatory cytokines and suppressed the degradation of Aβ plaque, leading to neuronal injury and AD progression (Wendimu and Hooks 2022). These reports indicate that activated microglia exhibit conflicting functions. Activated microglia were previously classified into proinflammatory (M1) and anti-inflammatory (M2) phenotypes, but this classification is insufficient for a complete understanding of these conflicting functions (Hashemiaghdam and Mroczek 2020). The recent development of a novel technique, which is single-cell RNA sequencing, showed that the presence of diverse phenotypes of microglia influences the pathogenesis of AD in the brain (Stratoulias et al. 2019). In addition, specific microglia phenotypes such as plaque-associated microglia or disease-associated microglia reportedly affect the pathogenesis of AD (Hashemiaghdam and Mroczek 2020;Keren-Shaul et al. 2017;Stratoulias et al. 2019). Hence, analysis of the diversity of microglia phenotypes and identification of diseaseassociated microglia may promote a better understanding of how microglia are involved in the pathogenesis.
Activated microglia are detected in the brain tissue infected with WNV (Kobayashi et al. 2012;Stonedahl et al. 2020). However, the roles of activated microglia remain controversial because they may exhibit either neuroprotective or neuroinjury-related roles (Chhatbar and Prinz 2021;Ghoshal et al. 2007;Malmlov et al. 2019;Seitz et al. 2018). Therefore, analysis of the diversity of microglia phenotypes and identification of disease-associated microglia may help to understand the neuropathogenesis of WNV infection. In this study, the diversity of microglia phenotypes caused by WNV infection was examined by detection of various microglia markers. Furthermore, disease-associated microglia were investigated in WNV-infected mouse brain tissue.

Virus and mice
The WNV NY99 6-LP strain was passaged in Vero cells (Kobayashi et al. 2020). All experiments using WNV were performed at a biosafety level 3 (BSL-3) in accordance with institutional guidelines. C57BL/6Cr Slc mice were obtained from Sankyo Labo Service (Tokyo, Japan) at 5 weeks of age. The mice were anesthetized using sevoflurane and infected with WNV through intracranial inoculation of either 100 or 10 plaque-forming units (pfu) per mouse. The virus solution was administered into the left frontal lobe of the brain. The mice were observed and weighed daily. Brain samples were collected after euthanasia at 5 days post-inoculation (dpi) in mice inoculated with 100 pfu and at 1, 4, and 6 dpi in those inoculated with 10 pfu.
For fluorescence visualization, the sections that reacted with the primary antibody were incubated with Alexa Fluor 488-, Alexa Fluor 555-, or Alexa Fluor 647-conjugated secondary antibodies with DAPI or Hoechst 33342 at room temperature for 1 h. For fluorescence labeling of antibody, anti-TMEM119 rabbit monoclonal antibody was labeled with Zenon rabbit IgG labeling kits (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer's instructions. All images were observed under an LSM 700 confocal laser scanning microscope with ZEN software (Carl Zeiss, Jena, Germany) and quantitated using ImageJ software (https:// imagej. nih. gov/ ij).

Quantitative analysis
To quantify fluorescence intensity, area, number of microglia, and distance measurements, the images were magnified 40-fold, saved in TIFF format, and imported to ImageJ. The file was separated into channels using the split-color command, allowing use of only the detected microglia channel. Individual microglia were selected in the region of interest. The signal fluorescence expression and area of microglia were individually measured using the integrated density and area commands, respectively. The distance was calculated from the nucleus of the microglia signal to that of the viral antigen-positive cell with ZEN software (Carl Zeiss). The number of microglia was manually counted under a microscope.

Preparation of single-cell suspension and flow cytometry
The brains of mice inoculated with 100 pfu were collected after euthanasia at 5 dpi and minced using scissors. The minced tissue was processed into a cell suspension in Hank's balanced salt solution (Gibco, Thermo Fisher Scientific) and DNase I (100 µL/mL; Roche, Mannheim, Germany). The cell suspension was filtered through a 70-µm cell strainer (Greiner Bio-One, Milan, Italy). The single-cell suspension was counted on a hemocytometer and collected after myelin removal with 30% Percoll solution (GE Healthcare, Uppsala, Sweden). The number of cells was set at a concentration of 2 × 10 6 cells/100 µL. The single-cell suspensions were blocked using 1% BSA-PBS for at least 10 min. Subsequently, single-cell suspensions were incubated with an APC-conjugated anti-mouse CD11c monoclonal antibody (eBioscience, San Diego, CA, USA) and Alexa Fluor 488-conjugated anti-rabbit TMEM119 monoclonal antibody Fig. 1 Microglia activation in the brains of mice inoculated with WNV. a Morphological differences in microglia between control and WNV-infected brains. The sections from the brains of control or WNV-infected mice were stained with hematoxylin and eosin (HE) or immunostained with antibody to WNV antigen and general microglia markers (TMEM119, Iba1, or P2RY12). The black arrow indicates perivascular cuffing, and the black arrowheads indicate degenerated neuronal cells. Scale bar: 100 µm. b Comparison of TMEM119-positive area between control and WNV-infected brains. The bar graph was created by measuring TMEM119-positive areas in individual microglia in each group using ImageJ software. The results are shown as mean ± standard deviation. *p < 0.05 in independent t-tests. c Enlargement of TMEM119-positive microglia by WNV infection. The sections from the brains of control or WNV-infected mice were immunostained with antibodies for WNV antigen and TMEM119. Scale bar: 50 µm (Abcam) for 30 min along with Fixable Viability Stain 780 (BD Biosciences, San Jose, CA, USA) for dead cell staining. The stained cells were fixed with 4% PFA for 15 min and filtered through a 40-µm cell strainer (Greiner Bio-One). The data were collected using a BD FACSLyric flow cytometer (BD Biosciences) and analyzed using BD FACSuite software v. 1.4 (BD Biosciences). The gating strategy was performed to discriminate doublets based on forward scatter area and forward scatter height. The debris and erythrocytes were separated from single cells based on forward scatter area and side scatter area. Then, these stained cells were investigated for CD11c and TMEM119 expression after exclusion of dead cells.

Statistical analysis
All variables were tested for normality by the Shapiro-Wilk test, and their distribution was determined. Comparison of two variables was performed using the independent Student's t-test for normally distributed data or the Mann-Whitney U-test for nonnormally distributed data. A multiple independent group with time dependent was analyzed using the Kruskal-Wallis analysis of variance. p-values < 0.05 were considered statistically significant. All statistical tests were performed using JMP Pro 16.1.0 (SAS Institute, Cary, NC, USA).

WNV infection induces microglia activation in the brains of mice
After inoculation of WNV, the mice showed loss of body weight and lack of activity from 4 dpi. Histological analysis of brain tissue inoculated with WNV revealed evidence of nonsuppurative encephalitis. In the hematoxylin and eosin sections, neuronal cell degeneration and perivascular infiltration of immunocytes were observed in various regions of the brain at 5 dpi (Fig. 1a). WNV antigen-positive cells were detected mainly in the cerebral cortex (Fig. 1a). To detect microglia in the brain tissues inoculated with WNV, we performed immunohistochemical analyses using three general markers for microglia. Cells positive for TMEM119, Iba1, or P2RY12 were detected in control and WNV-infected brain tissue (Fig. 1a). Enlarged and branched cells with a positive signal for each marker were observed in infected tissue but not control tissue (Fig. 1a). The positive area of TMEM119 in infected brain tissue was significantly larger than that in controls (Fig. 1b). These cells were stained simultaneously to reveal the relationship between WNV-infected cells and microglia. Enlarged cells with TMEM119-positive signals were located close to the viral antigen-positive cells (Fig. 1c). These results demonstrate that microglia were activated and underwent morphological changes due to WNV infection.

A variety of microglia phenotypes were detected in WNV-infected brains
Activated microglia in neurodegenerative diseases are not homogeneous and contain diverse phenotypes (Jurga et al. 2020). To examine the various microglia activated by WNV infection, we detected phenotypes of microglia using several microglia markers. Positive signals for ARG1, a marker of M2 alternatively activated microglia (Franco and Fernández-Suárez 2015), and SIGLEC-H, a marker for microglia that are absent from CNS-associated macrophages  The sections were immunostained with antibodies to WNV antigen and general microglia markers (TMEM119, Iba1, or P2RY12). The yellow box indicates a strong signal of microglia markers, and the gray box indicates a weak signal of microglia. Scale bar: 20 µm. d Measurement of the distance between TMEM119positive cells and WNV-infected cells. The distance was measured from the center of the nucleus of microglia to that of the nearest viral antigen-positive cell ( μm ), using ZEN software. Scale bar: 50 µm. e Comparison of the distance between TMEM119-positive microglia and WNV-infected cells. The bar graph shows the average distance between cells strongly or weakly expressing TMEM119 and the WNV-infected cells. The results are shown as mean ± standard deviation. *p < 0.05 in independent t-tests (Konishi et al. 2017), were detected in both the control and infected tissue (Fig. 2b, c, respectively). Positive signals for CD11c, a marker for activated microglia and dendritic cells (Shen et al. 2022), were detected only in infected tissue (Fig. 2d). Like TMEM119-positive cells, ARG1-positive cells were observed as enlarged and branched cells in the infected brain tissue (Fig. 2b). SIGLEC-H-positive cells were observed as enlarged rod shape in infected brain tissue (Fig. 2c). CD11c-positive cells in infected tissue were observed as small cells with branching processes (Fig. 2d). These findings suggest that WNV infection activates a variety of microglia in the brains of mice.

The microglia located close to the viral antigen-positive cells showed weak signal of general microglia markers
Next, we investigated the relationship between activated microglia and WNV-infected cells. The expression levels of TMEM119, Iba1, or P2RY12 in microglia located close to viral antigen-positive cells were weaker than those in microglia located far from viral antigen-positive cells (Fig. 3a, b, c, respectively). Then, we measured the distance from viral antigen-positive cells to each TMEM119-positive cell. The distance from cells that poorly expressed TMEM119 (weak signal) to the viral antigen-positive cells was significantly shorter than that from the cells that strongly expressed TMEM119 (strong signal) (Fig. 3d, e). These results suggest some interaction between microglia expressing weak signals of general microglia markers and WNV-infected cells.

Microglia weakly positive for TMEM119 in WNV-infected brain tissue are also positive for CD11c
In a recent study, microglia that were weakly positive for TMEM119 were detected in the brain tissues of patients with AD (Kenkhuis et al. 2022). Some of these cells were also positive for CD11c and were related to the pathogenesis of AD (Keren-Shaul et al. 2017). We examined the expression of CD11c in the microglia to characterize the microglia that were weakly positive for TMEM119 observed in WNVinfected brain tissue. A CD11c signal was detected from the microglia that were weakly positive for TMEM119 located close to viral antigen-positive cells (Fig. 4a). There were significantly more CD11c-positive cells in infected tissue than in control tissue at 4 or 6 dpi, and the number at 4 dpi was significantly higher than that at 1 dpi (Fig. 4b). Furthermore, the number of microglia that were both weakly positive for TMEM119 and positive for CD11c in infected tissue were analyzed by flow cytometry. The number was significantly higher in infected tissue than in controls (Fig. 4c, d). These results suggest that microglia that are weakly positive for TMEM119 and located close to WNV-infected cells were also positive for CD11c.

Discussion
In this study, we identified the various morphologies of microglia and differential expression of microglia markers in the brains of WNV-infected mice. In a variety of microglia phenotypes, CD11c-positive microglia were found close to WNVinfected cells. CD11c-positive microglia have been observed in brains affected by some neurodegenerative diseases; this is termed disease-associated microglia and may play a role in the pathogenesis (Keren-Shaul et al. 2017). Thus, the association between CD11c-positive cells and WNV-infected neurons suggests the importance of disease-associated microglia in the pathogenesis of West Nile encephalitis.
In a previous study, microglia were divided into M1 and M2 groups, similar to macrophages (Orihuela et al. 2016). However, this classification does not provide an adequate explanation of microglia differentiation (Han et al. 2019;Waltl and Kalinke 2022). Moreover, heterogeneous cellular phenotypes of microglia have been reported in several neurological diseases (Healy et al. 2022;Stratoulias et al. 2019;Tan et al. 2020). For instance, plaque-associated microglia and disease-associated microglia have been identified in the brain tissues of animal models of AD, and microglia with different expression levels of TMEM119 have been detected in the spinal cords of patients with amyotrophic lateral sclerosis and were considered disease-associated microglia (Hashemiaghdam and Mroczek 2020;Takahashi 2023). Consistent with these reports, we found many cells with different expression of microglia marker proteins in WNVinfected brain tissue. These findings indicate the diversity of microglia phenotypes caused by WNV infection. The elucidation of microglia phenotypes in some diseases is related to the development of prevention and treatment strategies (Stratoulias et al. 2019). In a study of AD, disease-associated microglia were detected using a singlecell transcriptomics technique (Keren-Shaul et al. 2017). A unique cluster of microglia that responded to AD plaque was found to be involved in the pathogenesis of not only AD but also other neurodegenerative diseases (Keren-Shaul et al. 2017;Satoh et al. 2019;Takahashi 2023). The diseaseassociated microglia concept has been revealed not only in slowly progressive diseases but also in acutely progressive conditions such as traumatic brain injury or viral infection (Spiteri et al. 2021;Todd et al. 2021). As a unique microglia type, CD11c-positive microglia were detected in the brain tissue of WNV-infected mice in our study. In previous studies, brain trauma was found to activate CD11c-positive microglia and increase the number of neuronal cells that underwent degeneration and death (Benmamar-Badel et al. 2020;Sato-Hashimoto et al. 2019). WNV infection mainly causes neuronal cell degeneration and death. Moreover, CD11c-positive microglia activated by WNV infection were located close to WNV-infected neurons in this study. Therefore, the neuronal cell degeneration and death caused by WNV infection seems to be associated with the induction of emergence of CD11cpositive microglia in the brain. Identification of the induction factor may elucidate the role of disease-associated microglia in the pathogenesis of WNV infection.
In this study, we observed microglia that were strongly or weakly positive for TMEM119, and some of the microglia weakly positive for TMEM119 were also positive for CD11c. This is consistent with previous studies that have shown that TMEM119 expression in disease-associated microglia is not high (Keren-Shaul et al. 2017;Satoh et al. 2019;Takahashi 2023). Investigation of TMEM119 expression may support detection of disease-associated microglia.
In summary, this work provides information on the microglia activated by WNV infection, including heterogeneous microglia phenotypes and disease-associated microglia. The disease-associated microglia may be associated with a protective role in the pathogenicity of WNV infection in the mouse brain because this microglia phenotype was previously reported to be associated with restricting the development of neurogenerative disease including AD (Keren-Shaul et al. 2017;Wlodarczyk et al. 2018). Analysis of the whole composition of the phenotype and the function or role of each phenotype will result in elucidation of the pathogenesis. Continued research in this field is expected to lead to the development of treatments for a variety of neurological diseases.