An Early Microglial Response Is Needed To Efficiently Control Herpes Simplex Virus Encephalitis

Microglia appear to be one of the principal regulators of neuroinflammation in the central nervous system (CNS). An increasing number of studies have demonstrated that the activation of microglia could result in either beneficial or detrimental effects in different CNS disorders. Hence, the role of microglia during herpes simplex virus encephalitis (HSE) has not been fully characterized. Using experimental mouse models, we showed that an early activation of the MCSF/CSF1R axis improved the outcome of the disease, possibly by inducing a proliferation of microglia. In contrast, depletion of microglia before HSV-1 infection worsened the prognosis of HSE. Thus, an early microglial response followed by sustained infiltration of monocytes and T cells into the brain seem to be key components for a better clinical outcome. These data suggest that microglia could be a potential target for immunomodulatory strategies combined with antiviral therapy to better control the outcome of this devastating disease.

ABSTRACT

The role of a signaling pathway through macrophage colony-stimulating factor (MCSF) and its receptor, macrophage colony-stimulating factor 1 receptor (CSF1R), during experimental herpes simplex virus 1 (HSV-1) encephalitis (HSE) was studied by two different approaches. First, we evaluated the effect of stimulation of the MCSF/CSF1R axis before infection. Exogenous MCSF (40 μg/kg of body weight intraperitoneally [i.p.]) was administered once daily to BALB/c mice on days 4 and 2 before intranasal infection with 2,500 PFU of HSV-1. MCSF treatment significantly increased mouse survival compared to saline (50% versus 10%; P = 0.0169). On day 6 postinfection (p.i.), brain viral titers were significantly decreased, whereas beta interferon (IFN-β) was significantly increased in mice treated with MCSF compared to mice treated with saline. The number of CD68⁺ (a phagocytosis marker) microglial cells was significantly increased in MCSF-treated mice compared to the saline-treated group. Secondly, we conditionally depleted CSF1R on microglial cells of CSF1R-loxP-CX3CR1-cre/ERT2 mice (in a C57BL/6 background) through induction with tamoxifen. The mice were then infected intranasally with 600,000 PFU of HSV-1. The survival rate of mice depleted of CSF1R (knockout [KO] mice) was significantly lower than that of wild-type (WT) mice (0% versus 67%). Brain viral titers and cytokine/chemokine levels were significantly higher in KO than in WT animals on day 6 p.i. Furthermore, increased infiltration of monocytes into the brains of WT mice was seen on day 6 p.i., but not in KO mice. Our results suggest that microglial cells are essential to control HSE at early stages of the disease and that the MCSF/CSF1R axis could be a therapeutic target to regulate their response to infection.

IMPORTANCE Microglia appear to be one of the principal regulators of neuroinflammation in the central nervous system (CNS). An increasing number of studies have demonstrated that the activation of microglia could result in either beneficial or detrimental effects in different CNS disorders. Hence, the role of microglia during herpes simplex virus encephalitis (HSE) has not been fully characterized. Using experimental mouse models, we showed that an early activation of the MCSF/CSF1R axis improved the outcome of the disease, possibly by inducing a proliferation of microglia. In contrast, depletion of microglia before HSV-1 infection worsened the prognosis of HSE. Thus, an early microglial response followed by sustained infiltration of monocytes and T cells into the brain seem to be key components for a better clinical outcome. These data suggest that microglia could be a potential target for immunomodulatory strategies combined with antiviral therapy to better control the outcome of this devastating disease.

INTRODUCTION

Herpes simplex virus 1 (HSV-1) is ubiquitous, and nearly 60% of the world population has antibodies against the virus (1–3). HSV-1 first infects innervating sensory neurons through an epithelial lesion and establishes a latent state in sensory ganglia (4, 5). HSV-1 usually causes asymptomatic infections or benign diseases, but it may also be responsible for severe pathologies, such as keratitis or encephalitis. Herpes simplex virus encephalitis (HSE) is the most common fatal sporadic viral encephalitis in industrialized countries. In the absence of treatment, the mortality rate of patients suffering from HSE is around 70% (6–8). Intravenous acyclovir therapy decreases this high mortality rate to 20%. However, most surviving patients still suffer from significant neurologic impairments. It is now commonly assumed that neuronal damage is caused not only by viral replication, but also by an exaggerated immune response in the brain (9, 10).

The mechanisms involved in the invasion of the central nervous system (CNS) by HSV-1 are not yet clearly understood. HSV-1 infiltrates the brain parenchyma and targets neurons and glial cells (11, 12). When cells of the CNS detect the presence of the virus, they induce the production of proinflammatory cytokines and chemokines (13). This first burst of inflammatory response is followed by the infiltration of peripheral immune cells into the CNS. This recruitment seems to be necessary to control viral replication and spread of the infection in the brain (14, 15). However, the prolonged, overactive inflammatory response that ensues may contribute to neuropathological sequelae (16, 17).

Microglia constitute the resident immune cell population of the CNS. Microglial cells are crucial in maintaining brain homeostasis (18). These glial cells are also involved in the protection of the brain parenchyma against pathogen invasions through different mechanisms, such as phagocytosis, the release of immune mediators, and antigen presentation via the major histocompatibility complex class II (19). The role of microglia in the context of HSV-1 infection of the CNS remains poorly explored.

Macrophage colony-stimulating factor (MCSF) is a hematopoietic cytokine that is expressed by a wide range of cells and tissues. In peripheral sites, the cytokine is involved in the development, proliferation, and maintenance of mononuclear phagocytes, such as monocytes, dendritic cells, and osteoclasts, through the activation of a tyrosine kinase family receptor, macrophage colony-stimulating factor 1 receptor (CSF1R), also known as cluster of differentiation 115 (CD115) (20). In the CNS, the MCSF/CSF1R axis is essential for the survival of microglia and also mediates their proliferation, phagocytic activity, and differentiation toward an anti-inflammatory phenotype (21–23). In this study, we first evaluated the effect of activation of the MCSF/CSF1R axis by exogenous MCSF administered before intranasal infection of mice with HSV-1 on the control of viral replication and inflammation in the brain. We then examined the implication of microglia during HSE in CSF1R-loxP-CX3CR1-cre/estrogen receptor 2 (ERT2) mice treated with tamoxifen to induce the conditional deletion of CSF1R in cells positive for C-X3-C motif chemokine receptor 1 (CX3CR1).

Here, we report that an early response of microglia is crucial for effective control of HSV-1 spread in the CNS and that its activation through the MCSF/CSF1R signaling pathway potentially ameliorates the outcome of HSE in murine models. Pretreatment with MCSF increases the number of microglial cells in the CNS of mice infected with HSV-1 and could enhance their phagocytic activity. We conclude that functional microglial surveillance is mandatory at the early stages of HSE for a better clinical outcome of the disease.

RESULTS

Pretreatment of mice with exogenous MCSF improves the prognosis of experimental HSE.

To study the effects of stimulation of the MCSF/CSF1R axis on the outcome of HSE, exogenous MCSF (40 μg/kg of body weight) or saline was injected intraperitoneally (i.p.) once daily into BALB/c mice on days 4 and 2 prior to intranasal infection with 2,500 PFU of HSV-1 strain H25 (Fig. 1A). Body weight losses were higher in mice treated with saline than in mice treated with MCSF on days 8, 9, and 10 postinfection (p.i.) (P < 0.01, P < 0.05, and P < 0.01, respectively) (Fig. 1B). The mortality rate was significantly higher in the mock-treated group (91.7%) than in the MCSF-treated mice (50.0%; P = 0.0169) (Fig. 1C). Taken together, our data indicate that pretreatment with MCSF improves disease outcome in mice infected intranasally with HSV-1.

FIG 1.

Early activation of the MCSF/CSF1R axis improves the survival rate and control of viral replication during HSE. (A) Four- to 6-week-old BALB/c male mice (n = 12 mice per group) received i.p. injections of MCSF solution (40 μg/kg) or saline (vehicle) once daily on days 4 and 2 before intranasal infection with 2,500 PFU of HSV-1 strain H25 in 20 μl minimal essential medium. (B) Percent body weight gain or loss of MCSF- and saline-treated mice. The numbers correspond to the numbers of dead mice on the indicated days for each group. (C) Survival curves of MCSF- and saline-treated mice. Survival rates were analyzed using a log-rank (Mantel-Cox) test. (D and E) Viral titers in homogenates of nasal turbinates (D) and the brain (E) were measured by a standard plaque assay on Vero cells on days 4 and 6 postinfection. The results are reported as PFU per milligram of nasal turbinates or brain homogenates and represent the means ± SEM for 3 to 5 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01.

Stimulation of the MCSF/CSF1R axis results in better control of HSV-1 replication in the brain during HSE.

The decreased mortality rates seen in MCSF-treated mice compared to the saline-treated group could result from reduced viral replication at the primary site of infection (nasal epithelium) or in the brain. We thus determined the infectious viral titers in tissue homogenates on days 4 and 6 p.i. for both groups of mice by a standard plaque assay on Vero cells. Viral titers were slightly lower in nasal turbinates (Fig. 1D) and brains (Fig. 1E) of mice treated with MCSF than in those of the saline-treated group on day 4 p.i. Interestingly, on day 6 p.i., brain viral titers were significantly reduced in MCSF-treated mice compared to control animals (P = 0.046) (Fig. 1E). More intense labeling was also detected on brain sections of the saline-treated group than in MCSF-treated mice on day 6 p.i., using an antibody directed against glycoprotein D (gD) of HSV-1 (data not shown). HSV-1-labeled proteins were mainly distributed in the olfactory bulbs, hippocampus, and hypothalamic regions. These data suggest that early activation of the MCSF/CSF1R axis results in better control of HSV-1 replication in the CNS.

MCSF stimulation increases the number of microglial cells at the early stages of HSE.

We first investigated by flow cytometry analyses whether exogenous MCSF administration could modulate the number of microglial cells and their phagocytic phenotype. The numbers of CD45^low/int CD11b⁺ CD115⁺ microglia were similar in MCSF and saline-treated groups prior to infection (Fig. 2A). However, the number of CD115⁺ microglial cells was 2 times higher in the brains of MCSF-treated mice (P = 0.048) on day 4 p.i., whereas it remained unchanged in the saline-treated group. The numbers of microglia in MCSF-treated mice then decreased on day 6 p.i., whereas an increase was seen in the saline-treated group. We observed an increase in the population of CD68⁺ microglia in MCSF-treated mice on day 6 compared to day 0 (P = 0.0001) and day 4 (P = 0.001) p.i. On day 6 p.i., the numbers of CD68⁺ microglial cells were 2 times higher (P = 0.05) in MCSF-treated mice than in the saline-treated group (Fig. 2B). To better characterize the implications of microglia, we then examined their interaction with HSV-1⁺ cells on brain sections of MCSF-treated mice on day 6 p.i. by immunofluorescence analyses. We observed that transmembrane protein 119-positive (Tmem-119⁺) microglial ramifications established a direct contact with HSV-1⁺ cells to form phagocytic engulfments (Fig. 2C, left). Moreover, antibodies specific to HSV-1 proteins Tmem-119 and CD68 were found to be colocalized ((Fig. 2C, right). Altogether, these data suggest that MCSF stimulation triggers the proliferation of microglia and induces a phagocytic phenotype.

FIG 2.

MCSF stimulation increases the number of microglial cells and induces their phagocytic phenotype during HSE. Shown are the numbers and phagocytic phenotypes of microglia in brain homogenates of mice treated with MCSF (gray bars and triangles) or saline (white bars and circles) before intranasal infection with HSV-1 strain H25. (A) (Top) Absolute numbers of CD11b⁺ CD115⁺ cells in the CD45^low/int gate corresponding to microglia were analyzed on days 0, 4, and 6 p.i. (Bottom) Flow cytometry plots for brain homogenates of mice treated with saline (left) or MCSF (right) on day 4 p.i. (B) Absolute numbers of CD68⁺ microglial cells. The results represent the means ± SEM for 3 to 5 mice per group at each time point. Statistical analyses were performed using ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001. (C) Immunofluorescence analyses of brain sections of MCSF-treated mice on day 6 p.i. Tmem-119⁺ (white) CD68⁺ (red) microglia surround HSV-1⁺ cells (green) in the hypothalamus region. The arrow indicates colocalized signals for HSV-1, Tmem-119, and CD68. The red circle shows microglial ramifications that form phagocytic engulfments. Nuclear staining was done with DAPI.

MCSF stimulation increases the infiltration of peripheral immune cells in the CNS.

The better control of viral replication in the brains of MCSF-treated mice than in the brains of the saline-treated group could also result from differences in the recruitment of peripheral immune cells to the CNS. We therefore evaluated their levels in the blood and brain homogenates of both groups of mice prior to (day 0) and at selected days after HSV-1 infection. We used a panel of 8 specific antibodies (Table 1) together with a LIVE/DEAD marker to distinguish different immune cell subsets by flow cytometry analyses. There was no significant difference in the blood levels of immune cells between the two groups of mice on day 0 and after HSV-1 infection (data not shown). The absolute numbers of CD45^high cells (infiltrating leukocytes) in brain homogenates were similar in both groups on day 4 p.i. (Fig. 3A). The number of CD45^high cells was 2.2 times higher in MCSF-treated mice than in the saline-treated group on day 6 (P = 0.0001) and then decreased to be comparable in both groups on day 8 p.i. Lymphocyte antigen 6 complex, locus C-positive (Ly6C⁺) monocytes/macrophages started to infiltrate the CNS on day 4 p.i. (Fig. 3B). Their levels increased significantly in mice treated with MCSF compared to mice treated with saline on day 6 (P = 0.0001) and then decreased to be slightly lower in the MCSF group than in mock-treated animals on day 8 p.i. Lymphocyte antigen 6 complex, locus G-positive (Ly6G⁺) neutrophils also started to infiltrate the CNS of both groups of mice on day 4, reached peak level on day 6, and markedly decreased with no significant differences between groups on day 8 p.i. (Fig. 3C). On day 4 p.i., CD3ε⁺ T (Fig. 3D) and B220⁺ B (Fig. 3E) lymphocytes infiltrated the brains of HSV-1-infected animals. On day 6 p.i., the levels of infiltrating T and B cells were higher in MCSF-treated mice than in the saline-treated group, but the differences were not statistically significant. On day 8 p.i., the absolute numbers of infiltrating CD3ε⁺ T and B220⁺ B lymphocytes were, respectively, 2.5 times (P = 0.0218) and 1.4 times (nonsignificant) lower in MCSF-treated mice than in mock-treated animals. These data suggest that MCSF modulates the recruitment of monocytes/macrophages to the CNS during HSE, i.e., induces higher infiltration on day 6 p.i. followed by lower infiltration at a later time after infection (day 8 p.i.).

TABLE 1.

Antibodies used for flow cytometry and immunofluorescence analyses

Assay	Target	Host	Clone	Fluorochromea	Manufacturer
Flow cytometry	CD45	Rat	30-F11	APC-Cy7	BD Biosciences
	CD11b	Rat	M1/70	BUV737	BD Biosciences
	CD11c	Hamster	HL3	PE-Cy7	BD Biosciences
	CD3ε	Hamster	145-2C11	PE-CF594	BD Biosciences
	B220	Rat	RA3-6B2	PerCP-Cy5.5	BD Biosciences
	Ly6C	Rat	AL-21	BV605	BD Biosciences
	Ly6G	Rat	1A8	FITC	BD Biosciences
	CD115	Rat	AFS98	APC	eBioscience
	CD68	Rat	FA-11	BV421	BD Biosciences
Immunofluorescence	Tmem-119	Rabbit	28-03	Unconjugated	Abcam
	HSV-1/2	Goat	Polyclonal	No	Bio-Rad
	CD68	Rat	FA-11	No	Bio-Rad
	Goat IgG	Chicken	Polyclonal	Alexa Fluor 488	Invitrogen
	Rat IgG	Donkey	Polyclonal	Alexa Fluor 647	Invitrogen
	Rabbit IgG	Goat	Polyclonal	Alexa Fluor 546	Invitrogen

^aAPC, allophycocyanin; PE, phycoerythrin; PerCP, peridinin chlorophyll protein; FITC, fluorescein isothiocyanate.

FIG 3.

MCSF stimulation increases the infiltration of peripheral immune cells in the CNS during HSE. Infiltration of peripheral immune cell subsets in brain homogenates of MCSF-treated (gray bars and triangles) and saline-treated (white bars and circles) BALB/c mice was analyzed by flow cytometry prior to (day 0) and on days 4, 6, and 8 following intranasal infection with 2,500 PFU of HSV-1. Shown are absolute numbers of CD45^high infiltrating leukocytes (A), CD45^high CD11b⁺ Ly6C⁺ monocytes/macrophages (B), CD45^high CD11b⁺ Ly6G⁺ neutrophils (C), CD45^high CD11b⁺ CD3ε⁺ T lymphocytes (D), and CD45^high CD11b⁺ B220⁺ B lymphocytes (E). The results represent the means ± SEM of 3 to 5 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001.

MCSF stimulation increases IFN-α/β mRNA levels during HSE.

To evaluate the innate antiviral response against HSV-1 infection in the CNS, we quantified alpha/beta interferon (IFN-α/β) mRNA levels in brain homogenates by droplet digital PCR (ddPCR) on days 4 and 6 after intranasal challenge with HSV-1. On day 4 p.i., the levels of IFN-α (Fig. 4A) and IFN-β (Fig. 4B) mRNAs normalized to those of the housekeeping 18S ribosomal subunit were not statistically different between the two groups of mice. Interestingly, a significant increase in the normalized levels of IFN-α (P = 0.03) and IFN-β (P < 0.001) transcripts was observed on day 6 compared to day 4 p.i. in mice treated with MCSF. Moreover, MCSF-treated mice had significantly higher normalized levels of IFN-β (P = 0.017) transcripts than the saline-treated group on day 6 p.i. These data suggest that stimulation of the MCSF/CSF1R axis could allow better control of HSV-1 replication in the CNS by increasing innate antiviral IFN-α/β responses.

FIG 4.

MCSF stimulation increases IFN-α/β mRNA levels during HSE. IFN-α (A) and IFN-β (B) mRNAs were quantified in brain homogenates of MCSF-treated (gray bars and triangles) and saline-treated (white bars and circles) BALB/c mice by droplet digital PCR on days 4 and 6 following intranasal challenge with 2,500 PFU of HSV-1. The numbers of mRNA copies for each IFN were normalized to those of the housekeeping 18S ribosomal subunit. The results represent the means ± SEM for 3 or 4 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01.

MCSF administration modulates inflammatory responses in the CNS.

We also compared the levels of cytokines and chemokines in brain homogenates of MCSF- and saline-treated mice on days 4 and 6 after intranasal challenge with HSV-1. The levels of cytokines and chemokines were not significantly different between the two groups on day 4 p.i. (Table 2). The levels of some cytokines were significantly increased in MCSF-treated mice compared to the saline-treated group (interleukin 13 [IL-13], P = 0.0249; IFN-γ, P = 0.0079; and IL-10, P = 0.0053) on day 6 p.i. In accordance with the increased infiltration of monocytes into the brain, we also observed higher levels of C-C chemokine ligand 2 (CCL2) (nonsignificant) in the CNS of mice treated with MCSF than in the saline-treated group. These data suggest that MCSF modulates the inflammatory responses mediated by microglia and, to some extent, by immune cells that infiltrate the brain.

TABLE 2.

Cytokine and chemokine levels in brain homogenates of BALB/c mice pretreated with MCSF

Cytokine/chemokine	Level (pg/ml) (mean ± SEM)a
	Saline-treated mice		MCSF-treated mice
	Day 4 p.i.	Day 6 p.i.	Day 4 p.i.	Day 6 p.i.
IL-1β	1.69 ± 0.23	9.01 ± 1.93	1.34 ± 0.17	11.74 ± 3.81
IL-2	3.69 ± 0.22	4.33 ± 0.31	3.20 ± 0.23	5.65 ± 0.50
IL-3	0.99 ± 0.12	2.61 ± 0.40	0.34 ± 0.19	4.43 ± 0.75
IL-4	0.29 ± 0.09	1.95 ± 0.61	0.09 ± 0.06	3.15 ± 0.88
IL-6	63.09 ± 20.96	2,656.93 ± 932.98	35.07 ± 17.36	1,491.71 ± 386.08
IL-10	17.11 ± 0.25	229.79 ± 39.29	10.34 ± 1.14	423.17 ± 41.10
IL-12p40	100.88 ± 10.42	193.05 ± 11.21	98.29 ± 23.92	204.39 ± 19.65
IL-13	136.65 ± 6.45	173.18 ± 17.66	124.80 ± 8.10	252.19 ± 15.13
IL-17A	5.74 ± 0.35	7.30 ± 0.47	3.61 ± 0.25	8.44 ± 0.81
G-CSF	479.57 ± 146.77	27,839.72 ± 8,539.83	536.1 ± 203.14	23,188.87 ± 6,519.42
GM-CSFb	ND	14.69 ± 6.55	ND	26.18 ± 5.52
TNF-α	35.19 ± 5.64	141.32 ± 78.99	26.76 ± 2.55	56.28 ± 6.23
IFN-γ	50.87 ± 13.05	474.21 ± 42.95	26.25 ± 5.28	949.45 ± 146.59
CCL2	651.48 ± 43.06	22,625.19 ± 3,731.42	582.52 ± 138.79	32,107.95 ± 6,298.12
CCL3	33.33 ± 5.25	924.03 ± 105.27	20.81 ± 3.14	1,125.85 ± 227.83
CCL4	12.67 ± 1.41	149.49 ± 12.24	9.46 ± 0.83	210.73 ± 57.25

^aCytokine/chemokine levels were analyzed by two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results compared to the control group (treated with saline) are indicated by boldface. P < 0.01. ND, not detected.

^bGM-CSF, granulocyte-macrophage colony-stimulating factor.

Liposomal clodronate treatment does not alter survival rates and brain viral titers during HSE.

In order to evaluate whether the improved survival rates of MCSF-treated mice were associated with higher production of monocytes/macrophages by bone marrow and/or infiltration in the brain, we induced their depletion by intravenous injection of liposomal clodronate or empty liposomes (control group) into BALB/c mice on days 3 and 5 after intranasal challenge with 1,000 PFU of HSV-1 strain H25 (Fig. 5A). The mortality rate of mice treated with liposomal clodronate was only slightly reduced compared to that of control animals, and the difference was not statistically significant (P = 0.29) (Fig. 5B). Viral titers were slightly lower in nasal turbinates of the control group than in those of mice treated with liposomal clodronate on day 4 p.i. (Fig. 5C), whereas no significant difference was seen in the brain (Fig. 5D). Interestingly, on day 6 p.i., brain viral titers were similar in the two groups despite higher viral titers in nasal turbinates of mice treated with empty liposomes than in those of the liposomal-clodronate-treated group.

FIG 5.

Liposomal clodronate treatment did not alter the survival rates of BALB/c mice during HSE. (A) Six-week old BALB/c male mice were infected intranasally with 1,000 PFU of HSV-1 strain H25. Groups of mice (n = 12 per group) received 200 μl of liposomal clodronate or empty liposomes by intravenous injection in the tail vein once daily on days 3 and 5 postinfection. (B) Survival curves of mice treated with liposomal clodronate and empty liposomes (controls). Survival rates were analyzed using a log-rank (Mantel-Cox) test. (C and D) Viral titers in homogenates of nasal turbinates (C) and brains (D) were measured by a standard plaque assay on Vero cells on days 4 and 6 postinfection. The results are reported as PFU per milligram of nasal turbinates or brain homogenates and represent the means ± SEM for 3 or 4 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: **, P < 0.01.

We then evaluated, by flow cytometry analyses, the levels of different immune cells in the blood and brain homogenates of mice treated with empty liposomes or liposomal clodronate prior to infection (day 0) and on days 4 and 6 after infection with HSV-1. There was no significant difference in the levels of CD45^high leukocytes in the blood between the two groups prior to and following HSV-1 infection (Fig. 6A). On day 4 p.i., the levels of CD45^high CD11b⁺ CD115⁺ Ly6C⁺ monocytes in the blood of mice treated with liposomal clodronate were significantly reduced compared to those of the control group (10% versus 40%; P = 0.0308) (Fig. 6B). The levels significantly decreased in the blood of mice receiving empty liposomes on day 6 p.i., which could correspond to monocytes leaving the blood to infiltrate the CNS. The numbers of other immune cells, such as neutrophils and T and B cells, remained unchanged in both groups of mice prior to and after HSV-1 infection (data not shown), indicating that the depletion induced by liposomal clodronate was specific to monocytes. In brain homogenates, the numbers of microglial cells and infiltrating CD45^high leukocytes remained unchanged in both groups throughout the infection (data not shown). Ly6C⁺ monocytes/macrophages were the only cell population for which infiltration significantly increased in the brains of control mice during infection (Fig. 6C), whereas the increase was nonsignificant in mice treated with liposomal clodronate. The levels of the two monocyte subsets Ly6C^high (Fig. 6D) and Ly6C^low (Fig. 6E) were significantly lower in the brains of mice treated with liposomal clodronate than in those of the control group on days 4 (P = 0.0268) and 6 (P = 0.0439) p.i., respectively. Altogether, our data suggest that monocytes/macrophages that infiltrate the brain contribute to a lesser extent to the control of viral replication in the brain than microglia.

FIG 6.

Blood monocyte depletion by liposomal clodronate leads to reduced monocyte infiltration in the brain during HSE. Shown are flow cytometry analyses of immune cells in the blood and brain homogenates of BALB/c mice infected with 1,000 PFU of HSV-1 strain H25 and treated with liposomal clodronate (gray bars and triangles) or empty liposomes (white bars and circles) on days 3 and 5 p.i. (A) Ratios of CD45^high leukocytes per living cell in the blood. (B) Ratios of CD45^high CD11b⁺ Ly6C⁺ monocytes/macrophages per total leukocytes in the blood. (C to E) Ratio of infiltrating CD45^high leukocytes per leukocyte (C) and CD45^high CD11b⁺ Ly6C^high (D) and CD45^high CD11b⁺ Ly6C^low (E) monocytes/macrophages per total leukocytes in the brain. The results represent the means ± SEM for 3 or 4 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Microglial depletion decreases the survival rates of mice infected with HSV-1.

To study the effect of microglial depletion on the outcome of HSE, we induced a conditional deletion of MCSF receptor selectively on microglia, using CSF1R-loxP-CX3CR1-cre/ERT2 mice. Wild-type (WT) (CSF1R-loxP × C57BL/6) and knockout (KO) (CSF1R-loxP × CX3CR1-creERT2) mice received tamoxifen by gavage once daily for 4 days (Fig. 7A). Three days after the last tamoxifen treatment, WT and KO mice (in a C57BL/6 background) were infected intranasally with 600,000 PFU of HSV-1 strain H25. Both groups of mice gained body weight until day 6 p.i. (Fig. 7B). Thereafter, the body weight of KO mice markedly decreased whereas it remained stable for WT animals. All the mice with microglial depletion died from the infection compared to 33% in the WT group (P < 0.005) (Fig. 7C).

FIG 7.

Microglial depletion decreases survival rates and interferes with the control of viral replication in the brain. (A) Five- to 7-week-old WT and CSF1R KO mice (n = 12 mice/group) received 5 mg of tamoxifen once daily for 4 days by gavage. Three days later, the mice were infected intranasally with 600,000 PFU of HSV-1 strain H25 in 20 μl minimal essential medium. (B) Percent body weight gain or loss of WT (circles) and KO (triangles) mice. The numbers correspond to the number of dead mice on the indicated day for each group. (C) Survival curves of WT and KO mice. The survival curves were analyzed using a log-rank (Mantel-Cox) test. (D and E) Viral titers in homogenates of nasal turbinates (D) and the brain (E) were measured by a standard plaque assay on Vero cells on days 4 and 6 postinfection. The results are reported as PFU per milligram of nasal turbinates or brain homogenates and represent the means ± SEM for 3 or 4 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results compared to those for the WT group are indicated as follows: *, P < 0.05; ***, P < 0.001. N.D, not detectable.

Depletion of microglia results in loss of control of brain viral replication during HSE.

We evaluated whether the marked difference in the mortality rates seen between WT and KO mice could be due to less efficient control of viral replication at the site of infection and/or in the brain. Viral titers in nasal turbinates were not significantly different in the two groups on days 4 and 6 p.i. (Fig. 7D). In contrast, brain viral titers were higher in KO mice than in WT animals on days 4 (nonsignificant) and 6 (P = 0.035) p.i. (Fig. 7E), which correlates with the mortality rates. These data suggest that the lack of control of HSV-1 replication in the brains of KO mice is not due to an altered mucosal response mediated by monocytes/macrophages at the site of primary infection.

Microglial response is needed at the early stages of HSV-1 infection.

To better characterize our mouse model, we first evaluated the levels of monocytes in the blood and microglia in the brains of WT and KO mice 3 days after the last dose of tamoxifen (which corresponds to day 0) by flow cytometry analyses. Although monocytes expressed CX3CR1, no significant difference in the number of CD45^high CD11b⁺ CD115⁺ Ly6C⁺ monocytes was observed in the blood of the WT and KO groups (Fig. 8A). As expected, a significant decrease in the number of CD45^low/int CD11b⁺ CD115⁺ microglial cells was observed in the brains of KO mice compared to those of their WT counterparts (P = 0.0028) (Fig. 8B). We then confirmed by immunofluorescence analyses of brain sections that the number of Tmem-119⁺ cells was markedly decreased in KO compared to WT mice on day 0 (Fig. 8C). These data show that this mouse model allows specific depletion of microglia.

FIG 8.

Conditional depletion of CSF1R in CSF1R-loxP-CX3CR1-cre/ERT2 mice leads to specific depletion of microglia. Five- to 7-week-old WT (white bars and circles) and KO (gray bars and triangles) mice received 5 mg/kg tamoxifen (TAM) by gavage once daily for 4 days or were left untreated. (A) Percent CD45^high CD11b⁺ Ly6C⁺ CD115⁺ monocytes per total leukocytes in blood of untreated or TAM-treated animals 3 days after the last dose. n.s., nonsignificant. (B) Absolute numbers of CD45^low/int CD11b⁺ CD115⁺ microglia in brain homogenates of KO and WT mice infected with HSV-1 at different times postchallenge. The results are expressed as means ± SEM for 3 to 8 mice per group. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. (C) (Top) Flow cytometry plots for CD45^low/int CD11b⁺ CD115⁺ gate. CD115 (CSF1R) depletion was confirmed in the CNS of WT and KO animals 3 days after the last tamoxifen treatment (which corresponds to day 0). (Bottom) Representative confocal microscopy images of brain sections labeled with Tmem-119 antibody (green) and counterstained with DAPI (blue) that also showed a lack of microglia in the hypothalamic region of KO mice (right) compared to WT (left) animals 3 days after the last tamoxifen dose.

On day 4 after intranasal challenge with HSV-1, the numbers of CD45^low/int CD11b⁺ CD115⁺ microglia were 3.5 times higher in the CNS of WT animals than in that of KO mice (P = 0.0005) and decreased on day 6 p.i. to be similar in the two groups (Fig. 8B). These data suggest that efficient activation of microglia on day 4 p.i. is essential to control HSV-1 spread in the CNS.

Recruitment of peripheral immune cells to the CNS is slightly altered in microglia-depleted mice at early stages of HSE.

We next evaluated by flow cytometry analyses whether microglial depletion affects the infiltration of peripheral immune cells in the CNS during HSE. The numbers of CD45^high leukocytes in the brains of WT animals were 1.7 and 2.5 times higher than those in the brains of the KO group on days 4 and 6 p.i., respectively, but the differences were not significant (Fig. 9A). On day 4 p.i., the recruitment of Ly6C⁺ monocytes/macrophages (Fig. 9B), Ly6G⁺ neutrophils (Fig. 9C), CD3ε⁺ T lymphocytes (Fig. 9D), and B220⁺ B lymphocytes (Fig. 9E) in the brain slightly decreased (nonsignificantly) in KO mice compared to WT animals. Thereafter, the infiltration of Ly6C⁺ monocytes/macrophages and CD3ε⁺ T lymphocytes, respectively, increased by 3.5 and 2.8 times in the brains of WT animals compared to KO mice on day 6 p.i. In contrast, only a slight increase of all immune cell subsets was seen in the brains of KO mice. Our data suggest that microglia could be involved in the recruitment of immune cells to the CNS. The lack of microglial cells orchestrating the inflammatory response at early stages of HSE could result in a lower mobilization of leukocytes, such as monocytes and T cells, to the CNS of KO mice, which could contribute, to some extent, to the loss of control of viral replication.

FIG 9.

Mice depleted of microglia show reduced numbers of monocytes and T cells in the CNS during HSE. Flow cytometry analyses show the infiltration of peripheral immune cell subsets in brain homogenates of WT (white bars and circles) and KO (gray bars and triangles) mice prior to (day 0) and on days 4 and 6 following intranasal infection with 600,000 PFU of HSV-1 strain H25. Shown are absolute numbers of CD45^high infiltrating leukocytes (A), CD45^high CD11b⁺ Ly6C⁺ monocytes/macrophages (B), CD45^high CD11b⁺ Ly6G⁺ neutrophils (C), CD45^high CD11b⁺ CD3ε⁺ T lymphocytes (D), and CD45^high CD11b⁺ B220⁺ B lymphocytes (E). The results represent the means ± SEM for 3 or 4 mice per group at each time point. Statistical analyses were performed using one-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05; **, P < 0.01.

Depletion of microglia reduces early expression of IFN-α/β in the brain.

As type I IFN secretion is critical in the control of HSV-1 replication, we evaluated their expression levels in brain homogenates of WT and KO mice prior to (day 0) and on days 4 and 6 after HSV-1 infection. The normalized levels of IFN-α (nonsignificant) (Fig. 10A) and IFN-β (P = 0.0179) (Fig. 10B) mRNAs were higher in the brains of WT mice than in those of KO animals on day 4 p.i., despite high viral titers in the brains of KO mice. Moreover, the two groups had similar normalized levels of type I IFNs on day 6 p.i. Our findings suggest that the absence of microglia could reduce the secretion of antiviral IFN-I in the brain at early stages of HSE.

FIG 10.

Mice depleted of microglia show reduced levels of IFN-α/β mRNA in the brain during HSE. IFN-α (A) and IFN-β (B) mRNAs were quantified in brain homogenates of WT (white bars and circles) and KO (gray bars and triangles) mice by droplet digital PCR on days 4 and 6 following intranasal challenge with 600,000 PFU of HSV-1 strain H25. The numbers of mRNA copies for each IFN were normalized to those of the housekeeping 18S ribosomal subunit. The results represent the means ± SEM for 5 to 7 mice per group at each time point. Statistical analyses were performed using two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results are indicated as follows: *, P < 0.05.

Depletion of microglia results in exacerbated cytokine production in the CNS at later stages of HSE.

The levels of selected cytokines and chemokines were measured in supernatants of brains harvested prior to infection (day 0) and on days 4 and 6 p.i. The production of cytokines/chemokines was quite similar in the two groups at early stages of infection (i.e., day 4 p.i.) (Table 3). On day 6 p.i., a significant increase in the levels of most cytokines/chemokines (IL-1α, P = 0.011; IL-1β, P = 0.0016; IL-2, P = 0.00307; IL-3, P = 0.0418; IL-5, P = 0.0011; IL-10, P = 0.0004; IL-12p40, P = 0.0027; IL-13, P = 0.032; IL-17A, P = 0.0106; granulocyte colony-stimulating factor [G-CSF], P = 0.0091; IFN-γ, P = 0.002; CCL2, P = 0.0002; CCL3, P = 0.028; CCL4, P = 0.0017; and CCL5, P = 0.011) was observed in KO animals compared to WT mice, except for IL-4, IL-6, and tumor necrosis factor alpha (TNF-α). These data suggest that an uncontrolled immune response occurred in mice depleted of microglia and that it was associated with the increased brain viral titers seen in these mice on day 6 p.i.

TABLE 3.

Cytokine and chemokine levels in brain homogenates of WT or KO mice during HSE

Cytokine/ chemokine	Level (pg/ml) (mean ± SEM)a
	WT mice			KO mice
	Day 0	Day 4 p.i.	Day 6 p.i.	Day 0	Day 4 p.i.	Day 6 p.i.
IL-1α	4.03 ± 0.08	4.63 ± 0.39	4.60 ± 0.59	3.92 ± 0.77	5.23 ± 0.38	7.98 ± 0.83b
IL-1β	1.04 ± 0.02	0.99 ± 0.05	1.04 ± 0.12	1.07 ± 0.11	0.92 ± 0.05	1.86 ± 0.19b
IL-2	2.59 ± 0.07	3.09	2.95 ± 0.07	3.09 ± 0.45	2.66 ± 0.12	4.02 ± 0.19c
IL-3	0.57 ± 0.10	0.57 ± 0.10	0.52 ± 0.09	0.57 ± 0.30	0.61 ± 0.13	1.32 ± 0.15b
IL-4	0.20 ± 0.08	0.08 ± 0.05	0.23 ± 0.06	0.08 ± 0.05	0.08 ± 0.05	0.39 ± 0.06
IL-5	0.67 ± 0.11	1.11 ± 0.21	0.99 ± 0.11	111 ± 0.16	1.58 ± 0.21	2.72 ± 0.37c
IL-6	1.33 ± 0.09	4.64 ± 1.96	5.87 ± 2.07	1.17 ± 0.14	9.84 ± 2.87	152.30 ± 77.09
IL-10	7.98 ± 1.62	6.21 ± 0.79	13.93 ± 3.29	5.67 ± 1.90	5.29 ± 0.46	55.10 ± 10.49d
IL-12p40	199.24 ± 17.70	185.82 ± 37.58	236.91 ± 7.12	136.18 ± 8.38	161.12 ± 22.72	468.27 ± 67.39d
IL-13	107.81 ± 5.84	111.24 ± 3.31	111.11 ± 6.74	104.50 ± 3.43	131.91 ± 9.66	145.70 ± 8.91b
IL-17A	2.62 ± 0.31	3.12 ± 0.10	3.22 ± 0.13	2.52 ± 0.80	2.35 ± 0.05	5.89 ± 0.64c
G-CSF	5.93 ± 0.43	66.31 ± 28.16	38.65 ± 13.08	9.31 ± 2.20	224.22 ± 62.92	774.09 ± 283.07c
GM-CSF	ND	ND	ND	ND	ND	ND
TNF-α	50.82 ± 20.74	44.39 ± 3.88	49.01 ± 6.04	61.17 ± 16.09	25.65 ± 1.20	60.28 ± 7.42
IFN-γ	9.95 ± 1.01	10.18 ± 0.40	13.26 ± 2.90	10.42 ± 1.46	8.22 ± 0.50	96.84 ± 27.03d
CXCL1	24,157 ± 3,614	47,010 ± 11,832	42,370 ± 9,045	29,543 ± 8,520	48,393 ± 2,852	98,083 ± 10,686c
CCL2	177.01 ± 34.83	275.14 ± 64.03	395.50 ± 104.29	353.72 ± 216.92	1,202.87 ± 76.93	3,231.88 ± 676.54d
CCL3	15.74 ± 1.03	19.54 ± 5.72	21.97 ± 2.50	19.01 ± 4.89	36.14 ± 2.70	107.35 ± 39.72c
CCL4	4.29 ± 1.61	4.16 ± 0.41	5.76 ± 0.45	5.99 ± 0.44	4.86 ± 0.40	15.81 ± 2.67d
CCL5	12,293 ± 5,679	27,853 ± 9,422	114,830 ± 27,871	9,063 ± 1,041	29,747 ± 2,912	340,294 ± 87,384b

^aCytokine/chemokine levels were analyzed by two-way ANOVA with a Tukey HSD post hoc test. Statistically significant results compared to the WT mice are indicated by boldface. ND, not detected.

^bP < 0.05.

^cP < 0.01.

^dP < 0.001.

DISCUSSION

Although the roles of microglia and the MCSF/CSF1R signaling pathway in the brain have been investigated in different contexts (24–27), its implication in HSV-1 infection of the CNS has not yet been widely explored. In this study, we analyzed the role of early microglial activation during experimental HSE by following two different approaches.

First, we exploited the early activation of the axis through the administration of exogenous MCSF to BALB/c mice prior to intranasal challenge with HSV-1. We chose to study the effect of MCSF pretreatment in BALB/c mice that were susceptible to HSE to better disclose the drug benefit for the outcome of the disease. Our results showed that MCSF-treated mice exhibited higher survival rates and better control of brain viral titers than the saline-treated group, indicating that signaling through this receptor plays a beneficial role during HSE. A significant increase in microglial numbers was observed in MCSF-treated mice on day 4 p.i., suggesting that the MCSF/CSF1R axis could trigger their proliferative response (28). It had already been demonstrated that microglia have the ability to phagocytose infected cells and cell debris during viral infection of the CNS with West Nile virus (29). Phagocytosis was identified as one of the main defense mechanisms used by microglia against invasions of pathogens. In this context, the MCSF/CSF1R axis was shown to play an important role in the regulation of the phagocytic activity of microglial cells. Immunohistochemistry analyses revealed that Tmem-119⁺ and CD68⁺ microglia colocalized with HSV-1⁺ cells in infected regions of the CNS, indicating that microglia could be involved in viral clearance. A similar observation was previously described in mice infected with a neurotropic pseudorabies virus (PRV) (another subfamily Alphaherpesvirinae member), as well as in postmortem brain tissues obtained from patients with HSE (30). We also showed that the number of CD68⁺ microglia in mice treated with MCSF was significantly increased compared to those of the saline-treated group on day 6 p.i. This suggests that MCSF stimulation could increase the phagocytic activity of microglia and contribute to viral clearance and limit the spread of infection in the brain. However, further studies are needed to understand the mechanism by which MCSF modulates microglial phagocytic activity.

On the other hand, the infiltration of monocytes/macrophages significantly increased in the brains of MCSF-treated mice on day 6 p.i. The higher recruitment of monocytes/macrophages into the CNS was associated with a slight increase in CCL2 secretion in brain homogenates (31). Mice treated with MCSF also exhibited overexpression of IFN-I transcripts on day 6 p.i., which could potentially lead to better control of the infection. In addition to these mediators, the levels of IFN-γ, IL-10, and IL-13 were also higher in the brains of MCSF-treated mice than in those of the saline-treated group on day 6 p.i. During viral encephalitis, monocytes were shown to regulate the recruitment of effector T cells to the site of viral infection (32). In mice treated with MCSF, increased infiltration of monocytes/macrophages into the brain could thus be responsible for higher recruitment of T cells producing IFN-γ (32). IFN-γ could also play a critical role in viral clearance and partly contribute to the reduced viral spread (33, 34). MCSF is also known to induce a switch toward an anti-inflammatory profile in microglia and monocytes/macrophages, which produce increased levels of IL-10 (35–38). Microglia activated by IL-10 exhibit a reduction of the inflammatory response, together with improved phagocytic activity (39). We also observed elevated IL-13 production in the MCSF-treated group. IL-13 was suggested to be an inducer of activated microglial death, resulting in enhancement of neuronal survival (40, 41).

To differentiate between the contributions of monocytes/macrophages and microglia during HSE, we then depleted monocytes by treatment of BALB/c mice infected with HSV-1 with liposomal clodronate. Mice treated with liposomal clodronate exhibited a slightly reduced survival rate compared to controls. The drug treatment induced a significant reduction in the levels of Ly6C^high “inflammatory” monocytes and Ly6C^low “patrolling” monocytes in brain homogenates on days 4 and 6 p.i., respectively. However, this reduced infiltration of monocytes/macrophages into the CNS did not alter brain viral titers in treated animals compared to controls. We thus suggest that, at the early stages of infection, infiltrating monocytes have a lower impact on the outcome of HSE. Previous studies from our laboratory have already demonstrated that the mortality rates of mice deficient in C-C chemokine receptor 2 (CCR2) in hematopoietic cells were lower than those of mice deficient in CX3CR1 in resident cells of the CNS, suggesting a more important role of functional CX3CR1⁺ microglia than of infiltrating CCR2⁺ monocytes in the control of HSE (42, 43).

Secondly, we set up an experimental protocol in CSF1R-loxP-CX3CR1-cre/ERT2 mice to induce conditional depletion of CSF1R by tamoxifen treatment. After infection with HSV-1, the mortality rates were markedly increased in mice conditionally depleted of CSF1R compared to their WT counterparts, with no survivors on day 8 p.i., suggesting a crucial role of microglia in protecting against HSE. In line with our results, the pharmacological depletion of microglia using Plexxikon 5622 (PLX5622), which blocks the MCSF/CSF1R signaling pathway, was also shown to increase brain viral titers and mortality rates in mouse models of West Nile virus encephalitis, mouse hepatitis virus (a neurotropic murine coronavirus) encephalitis, and PRV encephalitis, suggesting that the initiation of the innate and adaptive immune responses requires functional microglia in the CNS (30, 44, 45). The viral titers in nasal turbinates of the WT and KO groups were similar, suggesting that the depletion of CSF1R did not alter the mucosal response mediated by monocytes/macrophages. These data are in accordance with unchanged monocyte levels in the blood prior to HSV-1 infection. In contrast, a significant increase in brain viral titers was observed in KO mice compared to WT animals, suggesting that functional microglia are needed for efficient control of viral replication. Another study also suggested that activated microglia constitute an early innate barrier in the olfactory bulb against vesicular stomatitis virus spread within the CNS (46). Higher brain viral titers caused by the absence of microglial cells resulted in an exacerbated inflammatory response. The decrease in levels of IFN-α/β transcripts on day 4 p.i. in the brains of KO mice suggests that microglia play a crucial role in initiating and orchestrating the innate antiviral response to control HSV-1 replication in the CNS, as has been shown by others (47). Most cytokines/chemokines were significantly increased on day 6 p.i. in the CNS of KO mice, suggesting an overzealous inflammatory response that could be detrimental.

In our study, the infiltration of monocytes/macrophages in the CNS was significantly higher in WT mice than in the KO group on day 6 p.i., as previously demonstrated after PRV infection of mice depleted of microglia with PLX5622 (30). In contrast, increased monocyte/macrophage infiltration was observed in the brains of microglia-depleted mice infected with coronavirus. These results suggest that infection by different viruses may differentially modulate the microglial inflammatory response to recruit peripheral immune cells to the CNS (45). Microglia could thus play a key role in the recruitment of monocytes and T cells to the brain, as shown with mice treated with MCSF. The number of microglial cells slightly increased in both groups of mice on day 4 p.i., whereas it was significantly higher in WT than in KO mice. At later stages of HSE (day 6 p.i.), the microglial cell population decreased in WT mice. Recent studies have suggested that overactivated microglia could die by apoptosis or autophagy mechanisms to limit the inflammatory response (48, 49). In contrast, there were no significant changes in the number of microglial cells in KO mice during HSV-1 infection. Altogether, these data suggest that the number of microglia could be modulated during HSE to limit the spread of HSV-1 in the CNS at early stages of infection.

This study has some limitations. First, we acknowledge that exogenous MCSF is not specific to microglia but also exerts some effects on monocytes that could have influenced HSV-1 infection at peripheral sites. Despite this limitation, we were able to conclude that the microglial response seems more important than that of monocytes/macrophages in controlling viral replication in the brain during early stages of the disease. Secondly, the conditional depletion of CSF1R by administration of tamoxifen for 4 days did not result in complete depletion of microglia. Despite this fact, a marked increase in brain viral titers and mortality rates was observed in mice depleted of microglia compared to WT animals after intranasal challenge with HSV-1.

Thus, we report that pretreatment of mice with MCSF led to a better outcome for experimental HSE through early activation of microglia. MCSF stimulated the proliferation of microglia and induced a phagocytic phenotype. We suggest that the improved survival rate of MCSF-treated mice results from the combination of increased control of viral replication, enhanced phagocytosis, and anti-inflammatory response in the brain.

Furthermore, we showed the importance of an early microglial response in controlling viral replication by using mice conditionally depleted of MCSF receptor. Higher survival rates in the presence of microglia proves their pivotal role in the antiviral response. Our hypothesis is further supported by the fact that increased numbers of microglial cells induced by MCSF stimulation could lead to an improved outcome during HSE. Altogether, our findings highlight the potential of the MCSF/CSF1R signaling pathway in microglia as a novel target to control HSV-1 replication and inflammatory response in the brain. This axis could be used to boost early microglial responses to better control infections. However, we cannot ignore the fact that the stimulation of the MCSF/CSF1R axis could be detrimental in some diseases, such as cancers (26). The role of microglia in the control of HSV-1 infection of the CNS and strategies that could trigger microglial proliferation at early stages of HSE should be further studied to develop immunomodulatory strategies that could be combined with antivirals for the treatment of the disease.

MATERIALS AND METHODS

Animals.

Four- to 6-week-old BALB/c male mice were purchased from Charles River Canada. WT (i.e., CSF1R-loxP × C57BL/6 mice) and KO (i.e., CSF1R-loxP × CX3CR1-cre/ERT2 mice) mouse colonies were established and maintained in a C57BL/6 background at our research center. The animals were acclimated to standard laboratory conditions for 1 week and housed 3 or 4 per cage. The targeted environmental conditions (temperature and humidity) in the housing room were 22° ± 3°C and 50% ± 20% relative humidity. The light/dark cycle was 12 h (on at 7:15 and off at 19:15). All animals were used in accordance with the Canadian Council on Animal Care guidelines, and the protocol was approved by the Animal Care Ethics Committee of Laval University (protocols no. 2013078 and 2017072).

Pretreatment of BALB/c mice with MCSF prior to HSV-1 infection.

Groups of BALB/c male mice (n = 12 per group) were injected i.p. with MCSF (R&D Systems) diluted in saline (40 μg/kg) or vehicle on days 4 and 2 prior to intranasal infection with 2,500 PFU of clinical HSV-1 strain H25 (50) in 20 μl of minimal essential medium (MEM) (Thermo Fisher Scientific) (Fig. 1A). The mice were monitored for signs of sickness and mortality as described below. Subsets of mice (n = 4 or 5 per group and per time point) were sacrificed prior to infection (day 0) and at selected times p.i. for the determination of viral titers in nasal turbinates and the brain, levels of immune cells in the blood, infiltration of peripheral immune cells into the brain, IFN-α and IFN-β mRNA transcripts, and cytokine/chemokine synthesis in brain homogenates, as well as for immunohistochemistry analyses of brain sections as described below.

Depletion of circulating monocytes with liposomal clodronate.

Six-week-old BALB/c male mice were infected intranasally with 1,000 PFU of HSV-1 strain H25 in 20 μl MEM. Groups of mice (n = 12 per group) received 200 μl of liposomal clodronate or empty liposomes in phosphate-buffered saline (PBS) (∼5 mg/ml) by intravenous injection in the tail vein once daily on days 3 and 5 p.i. (Fig. 5A). The mice were monitored for HSE-related signs and mortality as described below. Subsets of mice (n = 3 to 5 per group and per time point) were sacrificed prior to infection (day 0) and at selected times p.i. for the determination of viral titers in nasal turbinates and the brain, levels of immune cells in the blood, and infiltration of peripheral immune cells in the brain, as described below.

Infection of CSF1R-loxP-CX3CR1-cre/ERT2 mice with HSV-1.

CSF1R-loxP-CX3CR1-cre/ERT2 mice (51) have tamoxifen-inducible CRE recombinase activity leading to deletion of the CSF1R gene in CX3CR1⁺ cells. WT and KO male mice were treated with 5 mg of tamoxifen dissolved in 100 μl corn oil by gavage once daily for 4 days (Fig. 7A). Three days later, the mice were infected with 600,000 PFU of HSV-1 strain H25 in 20 μl MEM by the intranasal route. Subsets of WT and KO mice (n = 4 or 5 per group and per time point) were sacrificed prior to infection (day 0) and at selected times p.i. for the determination of viral titers in nasal turbinates and the brain, levels of immune cells in the blood, infiltration of peripheral immune cells in the brain, IFN-α/β mRNA transcripts, and cytokine/chemokine synthesis in brain homogenates, as well as for immunohistochemistry studies on brain sections, as described below.

Evaluation of clinical signs of HSE.

Mice were monitored for HSE-related signs, namely, ruffled fur, ocular swelling, shaking movements, swollen forehead, body weight loss, and mortality, twice daily for 14 days. Animals were sacrificed when body weight loss equal to or greater than 20% or a combination of two other obvious sickness signs was observed.

Infectious viral titer measurements.

Mice were anaesthetized by i.p. injection of a mixture of ketamine and xylazine (at doses of 90 mg/ml and 10 mg/ml, respectively) and sacrificed by intracardiac perfusion with cold 0.9% saline. Nasal turbinates and brains were harvested and homogenized in PBS containing a protease inhibitor cocktail (Complete Mini EDTA-free protease inhibitor cocktail tablets) and a phosphatase inhibitor cocktail (PhosStop phosphatase inhibitor cocktail tablets) (both from Roche Diagnostics). A standard plaque assay on African green monkey kidney (Vero) cells (ATCC CCL-81; American Type Culture Collection) was used to determine viral titers in homogenates of nasal turbinates and brains. The numbers of PFU were determined as described previously (42).

Flow cytometry analyses of blood samples.

Blood samples (∼150 μl) were withdrawn from the facial veins of mice. Samples were collected in EDTA-coated tubes (Thermo Fisher Scientific) to prevent coagulation. A 100-μl volume of blood sample was incubated with 5 ml of red blood cell lysis buffer 1× (BioLegend) at room temperature for 20 min. After a single washing step, the cell suspensions were incubated with BD Horizon Fixable Viability Stain 510 (FVS510; BD Biosciences), followed by a blocking step with purified rat anti-mouse CD16/CD32 (Mouse BD Fc Block 2.4G2; BD Biosciences), both on ice for 30 min. After a washing step, immune cells were labeled with an antibody pool (Table 1) at 4°C for 30 min. Then, the cells were washed once, permeabilized with an eBioscience Foxp3 transcription factor staining kit (Thermo Fisher Scientific), and incubated with CD115 antibody. Labeled cells were fixed, washed, and resuspended in Dulbecco’s PBS (DPBS) (Wisent Bio Products). Finally, a 50-μl volume of Precision Count beads (BioLegend) was added to each sample in order to obtain an absolute cell count. Flow cytometry analyses and data acquisition were performed using a BD SORP LSR II (BD Biosciences) and BD FACS Diva software, respectively.

Flow cytometry analyses of brain homogenates.

Mice were deeply anaesthetized as described above and perfused intracardially with ice-cold DPBS without Ca²⁺ and Mg²⁺. Their brains were harvested, homogenized, and incubated in DPBS containing Liberase TL, Nα-p-tosyl-l-lysine chloromethyl ketone hydrochloride (TLCK) (both from Sigma-Aldrich), and DNase I (Biorbyt) for 1 h at 37°C. The homogenates were then filtered through a 70-μm cell strainer (BD Biosciences). After a washing step, the cells were centrifuged at 600 × g for 40 min in a Percoll gradient (37% and 80%; GE Healthcare). The cell ring at the interphase was collected, centrifuged at 300 × g for 10 min, and washed twice with DPBS containing 2% fetal bovine serum (FBS) (Wisent). The cells were incubated with CD16/CD32 (Mouse BD Fc Block 2.4G2) and then with FVS510, both on ice for 30 min. The cells were incubated with a pool of antibodies (Table 1) for 30 min. In some experiments, cells were washed once, permeabilized with an eBioscience Foxp3 transcription factor staining kit, and incubated with CD115 and CD68 antibodies. The labeled cells were fixed, washed, and resuspended in DPBS. Finally, a 50-μl volume of Precision Count beads was added to each sample. Flow cytometry analyses and data acquisition were performed using a BD SORP LSR II and BD FACS Diva software, respectively.

Immunofluorescence studies.

Mice were sacrificed by intracardiac perfusion with cold 0.9% saline followed by a 4% paraformaldehyde solution (pH 7.4) at 4°C. Fixed brains were cut in 25-μm coronal sections on dry ice, using a microtome (Reichert-Jung, Cambridge Instruments Company). Free-floating sections were washed 3 times in potassium PBS (KPBS) for 15 min each time and incubated in KPBS containing 2% bovine serum albumin (BSA) and 1% Triton X-100 (both from Sigma-Aldrich) for 30 min. The sections were incubated with primary antibodies (Table 1) diluted (1:1, vol/vol) in the same buffer solution at 4°C overnight. Following incubation, the sections were rinsed 3 times in KPBS for 10 min each time, followed by incubation with fluorochrome-conjugated secondary antibodies at room temperature for 90 min. Finally, nuclear staining was performed using DAPI (4′,6-diamidino-2-phenylindole) (diluted to 0.0002%; Molecular Probes) for 10 min. The sections were then rinsed 3 times in KPBS for 10 min each time before being mounted onto SuperFrost slides (Fisher Scientific) and coverslipped with Fluoromount-G (Electron Microscopy Sciences). Confocal fluorescence microscopy images were captured using a Confocal Quorum WaveFX spinning-disk confocal microscope (Quorum Technologies) equipped with a Hamamatsu ImageEM camera (Hamamatsu Corporation). Images were acquired using Volocity 4 software (Perkin Elmer).

Quantification of IFN-α/β mRNAs by ddPCR.

Total RNA was extracted from 1 ml of TRIzol-brain homogenate mixtures (900 μl and 100 μl, respectively) using a Direct-zol RNA MiniPrep Plus kit (Zymo Research Corporation) as described in the manufacturer’s instructions. The total RNA concentrations in the extracts measured using a NanoDrop One microvolume spectrophotometer (Thermo Scientific) were adjusted to 80 ng per total reaction mixture (30 μl) to generate cDNAs using random hexamers and SuperScript IV polymerase (both from Invitrogen). The reverse transcription (RT) reaction was performed as described in the manufacturer’s instructions. For the PCR step, a 5-μl volume of cDNAs was added to 20 μl of reaction mixture (QX200 ddPCR EvaGreen Supermix; Bio-Rad Laboratories) containing primers specific to IFN-α, IFN-β, and 18S ribosomal subunit transcripts (available upon request). PCRs were performed in droplets produced with the QX200 droplet generator (Bio-Rad Laboratories). Droplet-partitioned samples were then transferred to a 96-well plate, sealed, and cycled in a C1000 deep-well thermocycler (Bio-Rad Laboratories) under the following cycling protocol: 95°C for 5 min (DNA polymerase activation), 40 cycles at 95°C for 30 s, and 57.8°C for 1 min; post-cycling steps at 4°C for 5 min and then at 98°C for 10 min (enzyme inactivation); and an infinite hold at 12°C. The plate was then read in the FAM (6-carboxyfluorescein) channel of the QX200 droplet reader (Bio-Rad Laboratories). The data were analyzed using QuantaSoft software (version 1.7.4; Bio-Rad Laboratories). The numbers of mRNA copies of IFN-α/β were normalized to those of the housekeeping 18S ribosomal subunit in each sample.

Cytokine and chemokine level measurements.

Brain homogenates were centrifuged at 10,000 × g for 10 min at 4°C. Levels of cytokines and chemokines were determined in the supernatants by magnetic-bead-based immunoassays using the Bio-Plex mouse cytokine group I plex assay (Bio-Rad Laboratories) according to the manufacturer’s instructions. Data were analyzed using a Bio-Plex system equipped with Bio-Plex Manager software v6.0.

Statistical analyses.

All statistical analyses were performed using GraphPad Prism software v8 (GraphPad Software) after excluding outliers by Grubb’s test. A P value of <0.05 was considered statistically significant. All data are presented as means ± standard errors of the mean (SEM). Survival rates were analyzed by log-rank Mantel-Cox tests. All other data were analyzed by two-way analysis of variance (ANOVA), followed by a Tukey honestly significant difference (HSD) post hoc test.