Host Immune Responses in the Central Nervous System during Fungal Infections

SUMMARY

Fungal infections in the central nervous system (CNS) cause high morbidity and mortality. The frequency of CNS mycosis has increased over the last two decades as more individuals go through immunocompromised conditions for various reasons. Nevertheless, options for clinical interventions for CNS mycosis are still limited. Thus, there is an urgent need to understand the host-pathogen interaction mechanisms in CNS mycosis for developing novel treatments. Although the CNS has been regarded as an immune-privileged site, recent studies demonstrate the critical involvement of immune responses elicited by CNS-resident and CNS-infiltrated cells during fungal infections. In this review, we discuss mechanisms of fungal invasion in the CNS, fungal pathogen detection by CNS-resident cells (microglia, astrocytes, oligodendrocytes, neurons), roles of CNS-infiltrated leukocytes, and host immune responses. We consider that understanding host immune responses in the CNS is crucial for endeavors to develop treatments for CNS mycosis.

1 |. INTRODUCTION

Incidences of invasive opportunistic fungal infections have risen dramatically over the last two decades and are expected to continuously rise because of increasing numbers of individuals with impaired immunity due to HIV/AIDS, genetic disorders, aging, and immunosuppressive therapies. Accordingly, clinical cases of fungal infections in the CNS are also increasing. CNS mycoses are particularly severe due to limited treatment options, causing high morbidity and mortality. Over 5 million fungal species are estimated to exist, and about 70,000 fungal species have been described¹. Of these, only a few hundred are known to cause disease in humans². A small number of fungi are known to have neurotropism, and such fungal genera include but are not limited to Cryptococcus, Candida, Aspergillus, Mucor, Rhizopus, Blastomyces, Coccidioides, and Histoplasma. Even with this relatively limited variety of known fungal pathogens, CNS mycosis is difficult to treat and causes devastating outcomes to patients.

When fungal clearance fails at the initial infection site, neurotropic fungi disseminate and enter the CNS through the bloodstream³. Although the blood-brain barrier (BBB) restricts the entrance of fungi into the CNS parenchyma, fungi use various tactics to breach the barrier. The host response will then depend on CNS-resident cells to recognize the presence of fungi that have infiltrated through the BBB. Host CNS-resident cells indeed express a wide range of pattern-recognition receptors (PRRs). Yet, fungi also use stealthy approaches to evade host detection. Even when detection of invading fungi is successful, immune responses in the CNS require tight and masterful regulation to limit collateral damages to CNS-resident cells. Concomitantly, leukocytes from the circulation are also recruited to the infected CNS, further challenging the fine-tuning of host immune responses. This review intends to highlight the urgency of basic and mechanistic understanding of CNS immune response in the context of fungal infections.

2 |. FUNGAL PATHOGENS THAT INFECT THE CNS

Candida, Cryptococcus, and Aspergillus are the three major fungal genera that typically cause fungal infections and disseminate to the CNS^3–6. However, other fungi can also infect the CNS. These include non-Aspergillus molds, including fungi in the Mucorales order, and dimorphic fungi, such as Blastomyces, Coccidioides, and Histoplasma^3,6,7.

2.1 |. Candida species

Candida is a genus of yeast fungi within the phylum Ascomycota and is the most common infectious fungi. Globally, it is estimated that there are ~750,000 cases of invasive candidiasis per year, with up to 40% mortality, depending on patients’ predisposing factors⁵. Candida species can be harmless commensals located on mucosal surfaces, mainly in the gastrointestinal, respiratory, reproductive tracts, and the skin. Yet, they can be pathogenic when the mucosal or skin barriers are breached or compromised. Candida shows morphologies of yeast, pseudohyphae, or true hyphae during infection⁸. The majority of Candida-related CNS infections are caused by hematogenous spread of Candida albicans and cause meningoencephalitis⁴. In addition to disseminated Candida infection (candidemia), Candida CNS infections can be caused by neurosurgery of contaminated surgical devices⁷. Less commonly, Candida infection can cause chronic meningitis, brain abscesses, vasculitis with cerebral infarctions, spinal infections, ventriculitis, and mycotic aneurysms.

2.2 |. Cryptococcus species

Cryptococcus is a genus of Basidiomycota fungi. They are yeast-like fungi but can also exhibit a hyphal state. However, they are not traditionally thought to be dimorphic since hyphal production typically occurs during the mating process. Cryptococcus yeasts are surrounded by a thick “capsule” consisting of polysaccharides, such as glucuronoxylomannan (GXM)⁹. The capsule enables Cryptococcus to evade immune recognition¹⁰. Currently, the global incidence of cryptococcal meningitis is about 223,100 cases per year and causes 15% of AIDS-related deaths globally^5,11. Cryptococcus neoformans, an opportunistic fungus, is the most common species among the Cryptococcus genera. C. gattii is an endemic species, which infects immunocompetent individuals. An outbreak of C. gattii was reported two decades ago in the US Pacific Northwest¹². Meningitis and meningoencephalitis are typical features of CNS cryptococcosis, most often caused by C. neoformans, and less commonly by C. gattii. CNS cryptococcosis often arises from a primary lung infection. Cryptococcus infection is a common complication of untreated HIV infection¹³, which can cause immune reconstitution inflammatory syndrome (IRIS) upon immune reconstitution by HIV antiretroviral therapy^14,15.

2.3 |. Aspergillus species

Aspergillus fungi belong to Ascomycota, like Candida, but Aspergillus are conidial fungi, which go through asexual reproduction by forming conidia. Because Aspergillus conidia are airborne; lung infection through inhalation is the most common etiology of the infection¹⁶. Invasive aspergillosis is severe and aggressive. It is estimated that as many as 10 million individuals are at risk for invasive aspergillosis globally, with more than 300,000 cases reported yearly with an associated mortality range of 30–80%^5,17. Aspergillus fumigatus typically causes CNS infections through hematogenous spread from the primary site of infection, mainly the lung, or from a contiguous anatomical site such as the paranasal sinuses⁴. Although Aspergillus meningitis is relatively rare, cases are found more frequently in immunocompetent than immunocompromised patients¹⁸.

2.4 |. Mucorales order

Mucorales, including Rhizopus and Mucor, are mold fungi that belong to the phylum Mucormycota, considered to be phylogenetically more primitive than Basidiomycota and Ascomycota. Globally, the prevalence of mucormycosis is greater than 10,000 cases⁵ with mortality rates of up to 95%¹⁹. The most common clinical presentation is rhino-orbital cerebral mucormycosis (ROCM), which are mainly caused by Rhizopus oryzae²⁰. Although mucormycosis is relatively rare, immunocompromised individuals, as in hospital outbreaks, and diabetic patients are at high risk²⁰. COVID-19-associated mucormycosis recently reported from India also raised a grave concern²¹. Routes of infection are inhalation, open wounds, and ingestion of contaminated food²². Cerebral mucormycosis is considered to be the most aggressive CNS fungal infection and uniformly fatal⁷. Nevertheless, how our immune system responds to Mucorales infections is still largely unknown.

2.5 |. Other neurotropic fungi (including dimorphic fungi)

Despite being less common than the fungi described above, other fungal pathogens can infect the CNS. Depending on the growth conditions, they include dimorphic endemic fungi that grow as yeast or as mold²³. The group of fungi, including Blastomyces dermatitidis, Coccidioides spp., and Histoplasma capsulatum, can cause meningoencephalitis, hydrocephalus, brain or epidural abscesses, and spinal cord lesions⁶. Other groups of fungi that could infect the CNS are molds other than Aspergillus and Mucorales (including Fusarium, Scedosporium species, Dematiaceous molds)⁷ and yeast fungi other than Candida and Cryptococcus (including Sporothrix)²⁴.

3 |. MECHANISMS OF CNS FUNGAL INVASION

Fungal infiltration into the CNS parenchyma predominantly occurs when host immune responses are inadequate in clearing pathogens at the primary infection site. To establish CNS infection, fungi need to penetrate the BBB, which consists of four different cell types: endothelial cells, pericytes, astrocytes, and microglia. The BBB exists at the interface of blood vessels and interstitial fluid in the brain and spinal cord, limiting microorganisms from entering the CNS. Pathogenic fungi secrete various molecules that induce host cell death, damaging host barriers. Fungi can also “hitchhike” host immune cells, i.e., taking advantage of host cells as carriers, to invade the CNS.

In this section, we discuss mechanisms of fungal dissemination by the three major CNS-infecting fungal genera, Cryptococcus, Candida, and Aspergillus. There are three proposed models of fungal BBB penetration; transcellular migration (transcytosis), paracellular migration, and the Trojan Horse model. Reports have demonstrated that at least some species in Cryptococcus, Candida, and Aspergillus employ these dissemination methods to infiltrate the CNS.

3.1 |. Transcellular migration (transcytosis)

Transcellular migration (termed transcytosis) is achieved by two different mechanisms. One is absorptive-mediated transcytosis (AMT), which does not require ligand-receptor binding but utilizes charge interactions²⁵. The other mechanism is receptor-ligand mediated transcytosis (RMT), which requires specific binding between microbial ligand and a host cell receptor²⁵. Transcytosis is not limited to CNS entry. Fungi take advantage of host cell transcytosis to exit the initial site of infection and cross the barriers to enter the bloodstream or lymphatics.

The transversal crossing of C. neoformans through brain microvascular endothelial cells (BMECs) is the essential step for developing cryptococcal meningoencephalitis (CM). Adhesion and transcytosis of cryptococci through human BMECs are observed, even with an acapusular mutant of C. neoformans²⁶. As an example of RMT, hyaluronic acid (HA) on C. neoformans binds CD44 and ‘receptor of hyaluronan-mediated motility’ on human BMECs^27,28 (Fig. 1A). Indeed, CD44-deficient mice show a reduced fungal burden of C. neoformans in the brain²⁸. For cryptococci to attach to the brain endothelium, C. neoformans also secretes a metalloprotease, Mpr1²⁹. To assess the potency of Mpr1 in fungal dissemination, a study used Saccharomyces cerevisiae, which does not express Mpr1 and does not invade human BMECs³⁰. Artificial expression of Mpr1 in S. cerevisiae successfully allowed the fungus to pass through the brain endothelium²⁹. Because S. cerevisiae does not express HA, the result suggests that Mpr1 plays a critical role in transcytosis²⁹. Another study showed that host Annexin A2 and its synergistic partner S100A10 mediate C. neoformans transcytosis³¹. In addition, C. neoformans releases microvesicles (termed “virulence bags”), containing polysaccharides, lipids, and cytoplasmic proteins³². The microvesicle contents fuse with host cells, as observed in human BMEC culture and in vivo mouse model of CNS infection³². Such microvesicles are also considered to be involved in the adhesion of C. neoformans to host BMECs. Furthermore, when C. neoformans crosses human BMEC monolayers, the host cell actin cytoskeleton is rearranged without affecting the integrity of the endothelial barrier^26,33. Thus, without causing endothelial barrier disruption, C. neoformans can stealthily disseminate into the CNS without triggering host antifungal responses. Yet, a study suggested that C. neoformans ultimately induces endothelial cell necrosis³⁴.

Figure 1. Mechanisms of fungal brain invasion.

(A) Transcellular Migration: Interaction between a fungus and host barrier cells induces endocytosis of a fungus. Then, the fungus exits from host barrier cells and crosses the BBB. Shown is an example that HA in C. neoformans is detected by CD44 on BMEC to initiate receptor-ligand mediated transcytosis (RMT). (B) Paracellular Transport: Molecules secreted by fungi (e.g., proteases, urease) disrupt tight junctions between barrier cells to breach the BBB integrity. This allows the fungus to migrate across the endothelial layer into the brain parenchyma. (C) Trojan horse: Peripheral phagocytes uptake fungal spores or yeasts and migrate to the brain. Spores or yeasts are released from phagocytes once phagocytes migrate into the CNS. Created with BioRender.com

C. albicans express fungal invasins to achieve brain invasion. One invasin is Als3, which binds a heat shock protein, gp96, expressed specifically on the surface of brain endothelial cells and promotes Candida uptake by the host cells³⁵. Als3 also interacts with human cadherins to aid Candida endocytosis, as described in a study using human umbilical vein endothelial cells and human oral epithelial lines³⁶. Epidermal growth factor receptor (EGFR) and human epidermal growth factor receptor 2 (HER2) cooperatively induce endocytosis of C. albicans in an Als3-dependent manner³⁷, which complements the cadherin-dependent mechanism for endocytosis. These studies suggested Als3 as a critical molecule for Candida transcytosis. Independent of the yeast-mediated mechanism, the hyphal form of C. albicans also contributes to the loss of BBB integrity in disseminated candidasis³⁸. C. albicans can pass through the BBB by developing pseudo-hyphae, a virulence mechanism that does not elicit host cell morphological changes and enables the fungus to adhere, invade, and transcytose across human BMEC without affecting monolayer integrity³⁰.

Transcytosis of A. fumigatus across the BBB and infiltration into the CNS remains to be addressed. Yet, it is reported that sialic acid-specific lectin on A. fumigatus conidia adheres to fibrinogen or laminin on host cells³⁹. Additionally, galactosaminogalactan, a cell wall-associated polysaccharide of A. fumigatus, has been identified as an adhesin⁴⁰. Endocytosis of A. fumigatus conidia was identified in the A549 human lung epithelial cell line, human umbilical vein endothelial cells, and the J774 murine macrophage line⁴¹. Conidia internalization is promoted by fungal gliotoxin, which activates the actin cytoskeleton of host cells⁴². This mechanism is implicated in A. fumigatus transcytosis because gliotoxin alters the BBB integrity, at least ex vivo, without disrupting tight junction complexes⁴³. Once A. fumigatus gains access to the brain, conidia germinate and establish CNS infection.

3.2 |. Paracellular transport

In paracellular transport, microbes pass through barriers between epithelial cells by disrupting tight junctions without damaging host cells (Fig. 1B). Like transcytosis, paracellular CNS penetration also requires the attachment of fungi to the host cells before they are transferred.

C. neoformans employs paracellular transport^44,45, although paracellular transport is likely a minor route for migration across the BBB³⁴. Ureases, cryptococcal virulence factors⁴⁶, compromises barrier cell tight junction integrity⁴⁷. The approach is also used by C. albicans, which proteolytically degrades host E-cadherin to pass through epithelial cells, as reported in a study using oral epithelial cells⁴⁸. To target E-cadherin, C. neoformans uses host IL-33, which is highly expressed during pulmonary cryptococcosis⁴⁹. The study demonstrated that E-cadherin downregulation depends on IL-33⁴⁹.

Electron microscopy captured that A. fumigatus conidia, which are not engulfed by host immune cells, can form extracellular hyphae without lysing host cells⁵⁰.A. fumigatus produces proteases that disrupt the actin cytoskeleton in A549 cells⁵¹. A. fumigatus also produces mycotoxins, such as gliotoxin⁴², which are implicated in epithelium modifications during infection through paracellular transport, as well as transcytosis.

3.3 |. Trojan Horse Mechanism

Phagocytosis is a crucial mechanism in the clearance of pathogens. Yet paradoxically, phagocytic cells can become vehicles for fungal dissemination by serving as “Trojan Horses”^52,53 (Fig. 1C), because phagocytic host cells can pass through the BBB under certain conditions. As previously described, C. neoformans has a thick polysaccharide capsule, which can protect yeasts against reactive oxygen species (ROS) generated by host cells⁵⁴. The capsule enables C. neoformans to survive within the phagosome of human and mouse mononuclear phagocytes^55–57. Studies have demonstrated that monocytes containing C. neoformans are present in the perivascular space of the brain vasculature of infected mice^53,58 and transmigrate endothelial cell layers for C. neoformans to gain access to the brain⁵⁹. Depletion of monocytes or neutrophils with clodronate liposomes or Ly6G antibody, respectively, reduces cryptococcal burdens in brains of infected mice⁵³. Once C. neoformans crosses the BBB, the fungus exits macrophages⁶⁰ and neutrophils⁶¹ via non-lytic exocytosis without triggering host cell death-associated inflammation^60,61. Less is known about Candida and Aspergillus dissemination through the Trojan Horse mechanisms. However, a recent in vivo zebrafish study supported the Trojan Horse mechanism by demonstrating dissemination of C. albicans yeasts by using neutrophils as vehicles to invade the CNS⁶². A. fumigatus can also survive within phagocytes⁴¹, possibly due to the release of mycotoxins inhibiting macrophage oxidative burst⁶³. Thus, Candida and Aspergillus may use the Trojan Horse mechanism, but further investigation is necessary to determine if these fungi use this mechanism to infiltrate the CNS.

4 |. PATHOGEN DETECTION BY CNS-RESIDENT CELLS

In this section, we describe the capacity of CNS-resident cells to detect pathogens through PRRs. Although peripheral innate immune cells express PRRs, studies have shown that CNS-resident cells also express a variety of PRRs and respond to PRR signaling, including the production of chemokines and cytokines. PRRs recognize pathogen-associated molecular patterns (PAMPs), but PRRs also detect damage-associated molecular patterns (DAMPs), which are released by host cells that are injured, stressed, or died due to inflammation and tissue insults. This section focuses on PRR expression and responses by CNS-resident cells (Table 1) in fungal and other microbial infections. In some cases, we also described PRR signaling activated by endogenous molecules.

Table 1.

PRR Expression in CNS-Resident Cells

	Stimulators/Ligands Described	CNS-Resident Cell	References
Toll-like Receptors (TLRs)
TLR1 Dimerizes with TLR2	Triacyl lipopeptides α-synuclein	Microglia	72–75
		Astrocytes	75,165,166
		Neurons	202
TLR2	Peptidoglycan (PGN) Lipopeptides Lipoteichoic acid (LTA) Pam3CSK4 Zymosan	Microglia	72–75
		Astrocytes	172
		Oligodendrocytes	227,228
		Neurons	204,209,210
TLR3	poly(I:C)	Microglia	72–75,96,97
		Astrocytes	75,165,166
		Oligodendrocytes	228
		Neurons	206,211,212
TLR4	Lipopolysaccharide (LPS) Neuraminidase (NA) 4-hydroxynonenal (HNE)	Microglia	72–75,86,88,89
		Astrocytes	75,165,166,168,172
		Oligodendrocytes	229
		Neurons	204,210
TLR5	Flagellin	Microglia	72–75,106–109
		Astrocytes	75,165,166,172
		Neurons	106,211,212
TLR6 Dimerizes with TLR2	Diacyl lipopeptide	Microglia	72–75
		Astrocytes	75,165,166
		Neurons	202
TLR7	ssRNA Imidazoquinolinone R848 Imiquimod	Microglia	72–76,100,101
		Astrocytes	75,165,166,181
		Neurons	205
TLR8	ssRNA Imidazoquinolinone R848	Microglia	72–75
		Astrocytes	75,165,166
		Neurons	213
TLR9	CpG-DNA	Microglia	72–75,214
		Astrocytes	75,165,166,172,181
		Neurons	214
TLR10 (human)	Not determined	Microglia	74,75
TLR11 (murine)	Uropathogenic bacteria Profilin-like molecule from T. gondii	Microglia	72,73
		Astrocytes	75,165,166,184
		Neurons	184
TLR12 (murine)	Not determined	Microglia	72,73
		Astrocytes	75,165,166
TLR13 (murine)	Not determined	Microglia	72,73
		Astrocytes	75,165,166
		Neurons	73
C-type Lectin Receptors (CLRs)
DCIR	Not determined	Microglia	285
DC-SIGN	Mannose-containing glycoproteins Fucose Myelin/MOG	Microglia	116
Mincle	SAP130 α-mannose	Microglia	286
		Neurons	215,216
Mannose Binding Lectin (MBL)	Mannose Fucose N-acetylglucosamine Amyloid β peptides	Microglia	187
		Astrocytes	187
		Oligodendrocytes	187
		Neurons	186,187
Mannose Receptor (MR)	Mannose Fucose N-acetylglucosamine	Microglia	287
		Astrocytes	186
		Neurons	186
Dectin-1	β-glucans Zymosan	Microglia	113,124,127
NOD-like Receptors (NLRs)
NOD1	Diaminopimelic acid (DAP)	Microglia	137,138
		Astrocytes	188
		Neurons	188,217
NOD2	Muramyl dipeptide (MDP)	Microglia	137,138
		Astrocytes	188
NLRC4 (IPAF)	Palmitate	Microglia	147
		Astrocytes	146
		Neurons	219
NLRC5 (NOD4)	Not determined	Microglia	147
		Astrocytes	288
		Neurons	288
NLRX1 (NOD5)	Not determined	Neurons	221
NLRP1	Not determined	Microglia	144
		Neurons	218–220
NLRP2	ATP	Astrocytes	190
NLRP3	Lysophosphatidylcholine (LPC) Nigericin	Microglia	144–147
		Astrocytes	146
RIG-I-like Receptors (RLRs)
RIG-I	poly(I:C)	Microglia	148
		Astrocytes	195–197
		Neurons	208,224,225
MDA5	poly(I:C)	Microglia	148
		Astrocytes	195–198
		Neurons	208,224
AIM2-like Receptors (ALRs) and other PRRs
AIM2	Poly(dA:dT) dsDNA	Microglia	151,152,223
		Astrocytes	145,223
		Neurons	219,222,223
cGAS-STING	dsDNA	Microglia	154

4.1 ｜. Microglia

Microglia are CNS-resident macrophage-like cells and makeup 5–12% of brain cells in mice⁶⁴ and 0.5–16.6% of brain cells in humans⁶⁵. Microglia are involved in various stages of CNS biology, such as CNS development, homeostasis, and neuroinflammation. Microglia also serve as CNS sentinel cells to detect infections with highly developed phagocytic activity as macrophage-like cells. For the origin of microglia, monocytes from the circulation were previously considered to differentiate into microglia in the CNS. However, microglia are ontogenically distinct from other mononuclear phagocytes, including peripheral macrophages, and do not arise from hematopoietic stem cells in the bone marrow^66–68. A parabiosis approach clarified that microglia progenitors are not recruited from the circulation⁶⁸. Instead, in mice, microglia arise from the yolk sac at embryonic day 8.5–9.5 and migrate to the brain before the closure of the BBB^66,67. Microglia self-renew after being seeded in the brain and are maintained throughout life independently of the hematopoietic system^66,67. Unlike other macrophages, microglia resist high-dose irradiation^69,70. Yet, like peripheral macrophages, microglia express a broad range of PRRs, detect pathogens, participate in antigen presentation, and secrete various soluble molecules to communicate with other cells in the CNS⁷¹. As CNS-resident macrophage-like cells, microglia are the best-studied CNS-resident cells regarding PRRs (Table 1; Fig. 2A). However, a limited number of previous studies used in vivo microglia-specific mutant systems. Therefore, it is possible that some PRRs, which were previously identified in microglial tissue culture, might not be functional in vivo, although this issue also applies to CNS-resident cells other than microglia.

Figure 2. Pattern Recognition Receptors identified in CNS-resident cells.

CNS-resident cells express a variety of PRRs to detect DAMPs and PAMPs. (A) Microglia, as CNS-resident macrophage-like cells, express an extensive repertoire of PRRs, including TLRs, CLRs, NLRs, RLRs, and ALR. (B, C) Compared to microglia, less variety of PRRs, particularly CLRs, was reported in astrocytes (B) and neurons (C). (D) A limited variety of PRRs in oligodendrocytes (and oligodendrocyte progenitor cells) have been reported. Created with BioRender.com.

4.1.1: TLRs in microglia

Microglia constitutively express Toll-like receptors (TLRs), such as TLR1–9, TLR11–13 in mice^72,73 and TLR1–10 in humans^74,75, together with their adaptor proteins. Expression levels of various TLRs depend on the localization of microglia, age of hosts, and the condition of the CNS, i.e., setting of neurodegenerative diseases or infections^72,73,75,76. Microglial TLRs have been studied as inducers of proinflammatory responses against microbial pathogens. The functions of microglial TLRs have become a focus of intensive research. Together with C-type lectin receptors (CLRs), TLRs are critical for host fungal recognition. Studies on peripheral macrophages have revealed that fungal PAMPs are recognized by multiple TLRs (i.e., TLR2/1, TLR4, TLR3, TLR2/6, TLR7, and TLR9)⁷⁷. Among these TLRs, TLR2, 4, and 9 have been most intensively studied as detectors of fungal pathogens, including species of Candida, Cryptococcus, and Aspergillus^78–85, although further microglia-specific assessments of these TLRs will merit better understanding of CNS fungal infections.

Cell surface TLRs, including TLR4, are receptors that use Myeloid Differentiation primary response 88 (MyD88) as a key adaptor protein, which initiates activation of the transcription factors, mainly NFκB. TLR4 signaling induces the expression of proinflammatory cytokines⁷². In addition to well-known lipopolysaccharide (LPS), microglial TLR4 signaling is activated with neuraminidase (NA), a component of the envelope and membrane of some viruses and bacteria, respectively⁸⁶. Because C. albicans also expresses NA⁸⁷, microglial TLR4 may also be involved in Candida detection. An ex vivo study showed that TLR4 stimulation with LPS induces proinflammatory gene expression, such as Il1b, Il6, and Nos2, in primary microglia, although the same concentration of peptidoglycan, stimulating TLR2, induced expression of these genes to a slightly lesser extent⁷². More recent studies implicated enhanced microglial pyroptosis in response to TLR4 stimulation during spinal cord injury⁸⁸. TLR4 inhibition induced microglia to have an “M2” phenotype⁸⁹, characterized by the expression of Triggering Receptor Expressed on Myeloid Cells 2 (TREM2) and IL-10, which may promote angiogenesis and prevent neuronal death⁹⁰. TLR2 is also a cell surface TLR and forms a heterodimer with either TLR1 or TLR6. TLR2/1 and TLR2/6 detect triacyl lipopeptides and diacyl lipopeptides, respectively. TLR2/1 and TLR2/6 heterodimers on mouse macrophages detect GXM, a capsule component of C. neoformans⁹¹. However, A. fumigatus is detected only by TLR2/1 on human macrophages, despite mouse macrophages detecting A. fumigatus through both TLR2/1 and TLR2/6⁹². For detecting CNS endogenous molecules, microglial TLR2 detects neuron-derived α-synuclein to enhance neuroinflammation⁹³, where the involvement of TLR2/1 heterodimer is suggested⁹⁴. Nevertheless, preferential formation of TLR2/1 or TLR2/6 and specific outcomes of a heterodimer have been largely unexplored in microglia.

TLR3, TLR7, TLR8, and TLR9 are localized in the endosome, recognize nucleic acids, and are extensively studied in eliciting antiviral responses through type-I interferon (IFN-I) production⁹⁵. TLR3 detects double-stranded RNA (dsRNA) and depends on TRIF for downstream signaling as endosomal TLR4 does, while other endosomal TLRs require MyD88 for eliciting their signaling. Tlr3 mRNA is constitutively expressed in microglia⁹⁶. Primary microglia in culture respond to TLR3 stimulation with poly(I:C), by inducing gene expression, including Tlr3 and Ifnb, and secreting Tumor Necrosis Factor alpha (TNFα), IL-6, and IL-12p40⁹⁷. Microglia in Tlr3^−/− mice did not undergo microgliosis, an intensive reaction of microglia to pathogenic insults, after a poly(I:C) intracerebroventricular (i.c.v.) injection, while in wild-type mice, the treatment induced microgliosis⁹⁷. This study strongly suggested that TLR3 on microglia is biologically functional in vivo⁹⁷. Unexpectedly, Tlr3^−/− mice are more resistant than wild-type mice to encephalitic viral infections by West Nile virus (WNV) and a specific setting of Theiler’s murine encephalomyelitis virus infections^96,98. The detrimental role of TLR3 in WNV infection, at least, could be explained by TLR3-dependent inflammatory response, which allows CNS infiltration of WNV and neuronal injury⁹⁶. Although the role of microglial TLR3 is not known in fungal infections, dsRNA from A. fumigatus conidia activates lung epithelial cells through TLR3⁹⁹. Thus, TLR3 on microglia may also detect A. fumigatus conidia. However, it is unknown if microglial TLR3 is a beneficial or detrimental PRR in fungal infections.

Functions of other TLRs in microglia are still largely unexplored in fungal infection settings. Here, we briefly describe previous studies on roles of the microglial TLRs. The following studies were not performed in the context of fungal infections, but the findings might be extrapolated to possible outcomes when these TLRs are stimulated by fungi. TLR7 detects single-stranded RNA (ssRNA), and its expression has been observed in human⁷⁶ and mouse microglia¹⁰⁰ in numerous neurodegenerative diseases. In vivo administration of R848, a TLR7 agonist, induces acute sickness associated with CNS proinflammatory gene expression and microglia morphology changes¹⁰¹; thus, TLR7 is functional in microglia in vivo. TLR8 also recognizes ssRNA and is structurally related to TLR7. TLR8 is expressed in microglia⁷², but its function has yet to be addressed. TLR9 detects microbial and endogenous dsDNA. Previous studies suggested that TLR9 stimulation has substantial impacts on the CNS^102–105, but it is still unknown if and how microglial TLR9 is involved in fungal infections. Additionally, although it is clear that TLR9 is expressed and functional in microglia, experimental systems to clarify the involvement of microglia-specific TLR9, such as using TLR9 conditional knockout mice, will clarify the implication of TLR9 stimulation in microglia in vivo. Other microglial TLRs reported include TLR5, which detects bacterial flagellin. TLR5 expression has been described in human and mouse microglia^72,76, and its functional relevance has been explored in CNS disorders such as neuropathic pain, stroke, and Alzheimer’s disease (AD)^106–108. Ex vivo microglial TLR5 stimulation with flagellin results in the release of proinflammatory cytokines, promoting chemotactic microglial migration, and enhancement of phagocytosis¹⁰⁹. TLR10 is a human receptor, and its expression and function in microglia remain a subject for further research. TLR11–13 are murine TLRs, and their expression is elevated during murine neurocysticercosis, a parasitic infection that causes neurologic syndromes such as epileptic seizures^73,110,111, but their functions in microglia remain unknown.

4.1.2: CLRs in microglia

Microglia also highly express a variety of CLRs, expressed primarily by myeloid cells (Table 1; Fig. 2A). In peripheral immune cells, dectin-1, dectin-2, MCL/dectin-3, Mincle, mannose receptor (MR), and Dendritic Cell-Specific Intercellular adhesion molecule-3-Grabbing Non-integrin (DC-SIGN) have been studied as detectors of fungal pathogens¹¹². Among them, microglia express dectin-1, Mincle, MR, and DC-SIGN^113–116. Dectin-2 does not appear to be expressed at detectable levels in human or murine microglia¹¹⁷. Dectin-1, encoded by the Clec7a gene, was the first CLR identified in macrophages to detect fungi^118,119 through detection of β-glucans, cell wall components of fungi, some bacteria, and plants¹²⁰. Dectin-1 on macrophages collaborates with TLR2 and MR for heightened fungal sensing¹²¹. Although dectin-1 is not required for detecting C. neoformans¹²², it is crucial to recognize Candida and Aspergillus^118,123. Clec7a^−/− mice infected with C. albicans showed increased fungal loads in the brain¹²⁴. Thus, dectin-1 limits brain fungal loads in C. albicans infection but not in infections by C. glabrata, C. tropicalis, and C. parapsilosis¹²⁴, suggesting that the role of dectin-1 in Candida is specific to species. Interestingly, a study showed that β-glucan stimulation is not sufficient for substantial production of cytokines or chemokines by microglia, although ROS generation and phagocytosis are actively induced¹²⁵. In particular, the involvement of vav guanine nucleotide exchange factor 1 (Vav1) and phosphatidylinositol 3-kinase (PI3K) was identified in ROS generation by microglia¹²⁶.

Dectin-1 is a critical receptor in eliciting host responses against fungal infections. Therefore, this section is intended to discuss microglial dectin-1 in neuropathological settings other than fungal infections. Clarifying if dectin-1 signaling in microglia shares common outcomes and mechanisms between CNS mycosis and CNS sterile inflammation will merit further studies. Recent studies have revealed the involvement of dectin-1 on microglia in various sterile and non-homeostatic settings. Microglia, in mice at least, have no or very low expression of dectin-1 during homeostasis, but its expression is significantly upregulated during neurodevelopment and under CNS pathological conditions^{113,127–130} (more details described in our previous review article¹³¹), and its expression is tightly upregulated in the spinal cord of mice with experimental autoimmune encephalomyelitis (EAE)¹¹³ and in the brain of aged APP-PS1 mice with β-amyloid accumulation¹²⁷. Indeed, recent studies reported subsets of microglia, including microglial neurodegenerative phenotype (MGnD) and disease-associated microglia (DAM), highly express Clec7a in the brains of mice and humans in neuroinflammatory and neurodegenerative diseases^127–129. Dectin-1-expressing microglia are associated with β-amyloid plaques and diffuse neuritic dystrophy in the cortex of AD model mice¹²⁷. In addition, Clec7a upregulation was also identified in early postnatal proliferative region-associated microglia (PAM) during sterile CNS development¹³⁰. PAM exhibit amoeboid morphology, are metabolically active, and phagocytose newly formed oligodendrocytes¹³⁰. Thus, dectin-1 is highly expressed in microglia during pathological and neurodevelopmental settings independent of microbial infections¹³¹. Dectin-1 could function as an immunoregulatory receptor by detecting endogenous molecules in non-infectious settings^113,132, and the immunoregulatory role of dectin-1 might become simultaneously active with the canonical proinflammatory role of dectin-1 during CNS fungal infections.

DC-SIGN in peripheral antigen-presenting cells recognizes mannose-containing glycoproteins on the surface of pathogens, such as C. albicans, Mycobacterium tuberculosis, and Helicobacter pylori¹³³. Human microglia also express DC-SIGN¹¹⁶. Other studies demonstrated that DC-SIGN serves as a co-receptor for several viruses^134,135. Unstimulated microglia show little to no expression of DC-SIGN, but DC-SIGN is detected in primary human microglia stimulated ex vivo^116,136. DC-SIGN on dendritic cells (DCs) and microglia recognizes myelin, which is rich in carbohydrates and mediates internalization and endolysosomal routing of large myelin particles¹¹⁶. Costimulation of DC-SIGN and TLR4 enhances DC IL-10 production upon detecting myelin oligodendrocyte glycoprotein (MOG)¹¹⁶. The IL-10-producing immunoregulatory effect by DCs was abrogated when myelin had less fucosylated glycans present¹¹⁶. Thus, DC-SIGN may also play an immunoregulatory role in CNS fungal infection.

4.1.3: NLR in microglia

The nucleotide-binding oligomerization domain (NOD)-like receptors (NLRs) are intracellular sensors. The NOD subfamily of NLRs include NOD1 and NOD2, which recognize peptidoglycan motifs found mainly in bacteria. Expression of NOD1 and NOD2 was identified in microglia^137,138. In pulmonary aspergillosis, NOD1 is detrimental to hosts by limiting host defense against Aspergillus¹³⁹, but NOD2 is beneficial for host resistance¹⁴⁰. NOD2 is also critical for the inflammatory response against neurotropic bacteria^141,142.

The Nucleotide-binding oligomerization domain, Leucine-rich Repeat and Pyrin domain containing (NLRP) subfamily consists of intracellular sensors that mediate inflammasome activation by sensing a wide range of PAMPs and DAMPs. In particular, NLR Family Pyrin Domain Containing 3 (NLRP3) and NLR Family CARD Domain Containing 4 (NLRC4) inflammasomes detect Candida, Cryptococcus, Aspergillus, and Paracoccidioides (reviewed in¹⁴³). Microglia express components necessary for the formation of inflammasomes, including NLRP1, NLRP3, pro-caspase-1, and ASC¹⁴⁴. Expression of inflammasome components alone does not necessarily mean inflammasomes are activated or functional, because polymerization of inflammasome components is essential for inflammasome activation. For example, our lab recently reported that inflammasomes are not activated in microglia in the spinal cord during EAE¹⁴⁵, although expression of NLRP3 inflammasome components is detected in microglia. Due to the technical limitation in detecting active inflammasomes in the CNS, most previous studies stopped short of demonstrating in vivo or in situ inflammasome activation in microglia and other CNS cells. Instead, many studies described the expression of inflammasome components or inflammasome activation ex vivo. For example, the NLRP3 inflammasome is activated ex vivo in microglia in response to lysophosphatidylcholine (LPC), a molecule associated with neurodegeneration and demyelination¹⁴⁶. In a model of ischemia, NLRC5 is upregulated in microglia and directly binds to NLRP3 and NLRC4 in a nucleotide-binding domain-dependent manner to cooperatively drive microglial pyroptosis¹⁴⁷.

4.1.4: RLRs, and other PRRs in microglia

RIG-I-like receptors (RLRs) are cytosolic RNA sensors, which trigger the production of IFN-I. The RLR family is composed of three members: retinoic acid-inducible gene I (RIG-I), melanoma differentiation-associated protein 5 (MDA5), and laboratory of genetics and physiology 2 (LGP2). Isolated mouse microglia constitutively express RIG-I and MDA5 transcripts and proteins¹⁴⁸. Although RLRs have been extensively studied in the context of antiviral immunity, the involvement of RLRs in antifungal immunity has only recently been reported. One report using mouse macrophages and human PBMCs has shown that the yeast-to-hyphae transition of C. albicans induces expression of IFIH1, encoding for MDA5¹⁴⁹. The study also showed that MDA5-deficient splenocytes failed to recognize C. albicans hyphae, resulting in reduced IFNβ expression¹⁴⁹, suggesting recognition of C. albicans hyphae by MDA5. Furthermore, patients suffering from chronic mucocutaneous candidiasis have lower levels of IFIH1, suggesting the involvement of MDA5 in systemic candidiasis¹⁴⁹. Based on the constitutive expression of MDA5 in microglia¹⁴⁸, the involvement of microglial MDA5 in CNS candidiasis is expected.

AIM2-like receptors (ALRs) usually contain an N-terminal pyrin (PYD) domain, and one or two HIN domains. Absent In Melanoma 2 (AIM2) is a cytoplasmic DNA sensor, which forms an inflammasome upon detecting exogenous and endogenous dsDNA. Although the study focused on DCs in mice, A. fumigatus induces the formation of the AIM2 inflammasome complex, which also includes NLRP3 and caspase-8, in addition to canonical inflammasome components, ASC and caspase-1¹⁵⁰. With this mechanism, the AIM2 inflammasome protects hosts from pulmonary aspergillosis¹⁵⁰. Microglia express AIM2, and stimulation with poly(dA:dT) induces IL-1β and IL-18 release from microglia¹⁵¹, suggesting that the AIM2 inflammasome can be activated in microglia ex vivo. A more recent study showed that AIM2 in microglia limits EAE in an inflammasome-independent mechanism by negatively regulating cyclic GMP-AMP synthase (cGAS) and the DNA-dependent protein kinase (DNA-PK) pathways¹⁵². The study also demonstrated in an EAE setting that AIM2-deficient microglia induce the expression of genes related to immune responses against virus infections¹⁵². Thus, the study suggested that AIM2 in microglia negatively regulates inflammatory immune responses without forming an inflammasome. It is unknown if the inflammasome-independent AIM2 in microglia is functional in CNS fungal infections.

The cGAS- Stimulator of Interferon Genes (STING) pathway senses cytosolic dsDNA and induces IFN-I production by IRF3 activation. Although the study focused on human corneal epithelial cells (HCECs) and mouse corneas, cGAS-STING contributes to the progression of A. fumigatus keratitis¹⁵³. Yet, it is not known which source of DNA, host or A. fumigatus, stimulates the cGAS-STING pathway. Another study identified the cGAS-STING pathway activation in Iba1-positive cells in the brain of mice in the middle cerebral artery occlusion (MCAO) model¹⁵⁴, suggesting the possible presence of the cGAS-STING pathway in microglia. Using a microglia BV2 cell line, the study also demonstrated the activation of cGAS-STING to unspecified DAMPs¹⁵⁴. Taken together, it is plausible that cGAS-STING activation in microglia occurs during CNS infection, but this possibility remains to be addressed.

4.2 |. Astrocytes

Arising from primary neural stem cells, astrocytes are the most abundant cell population in the CNS and contribute to CNS functions in health and disease. During CNS homeostasis, astrocytes are critical cells for the formation and maintenance of the BBB; controlling extracellular fluid, ions, and neurotransmitters; providing energy substrates to neurons; modulating blood flow; and regulating drainage of interstitial fluid^155–158. Astrocytes also have roles in synapse development and plasticity¹⁵⁹, and interact with various cells in the CNS. Despite significantly fewer reports on pathogen sensing by astrocytes than microglia, previous studies demonstrated the involvement of astrocytes in shaping the immunological milieu upon microbial infections and neurodegeneration^113,160,161. In these pathological conditions, astrocytes undergo morphological and functional changes in a process known as reactive astrogliosis, in which the most common hallmarks include upregulation of glial fibrillary acidic protein (GFAP), the main component of astrocyte intermediate filaments and hypertrophy of astroglial processes¹⁶². During this process, astrocytes can form a glial scar around lesions during trauma or infection¹⁶³. Additionally, astrocytes produce cytokines, chemokines, and growth factors, although astrocytes express less variation of PRRs than microglia (Table 1; Fig. 2B). Technical challenges in astrocyte handling, such as obtaining a highly pure astrocyte population and their susceptibility in ex vivo environments, hinder functional analyses of astrocytes. Their functions during fungal infections remain largely unknown. At least, the formation of an astrocytic border around lesions has been demonstrated in several mouse models of fungal infections^28,164, although direct interactions between astrocytes’ PRRs and fungi have not been elucidated.

4.2.1: TLRs in astrocytes

Roles of astrocytic TLRs in fungal infections have yet to be explored, and limited information is available. In this subsection, we discuss previous findings, which provide knowledge to be considered when the roles of astrocytic TLRs in fungal infections are studied. The majority of TLR studies with astrocytes were performed ex vivo and focused on TLR expression or release of cytokines and chemokines by astrocytes in response to canonical TLR ligands. TLR expression under the homeostatic condition is reported in humans (TLR1–5, TLR9) and mice (TLR1–9 and TLR11–13) astrocytes^75,165,166 (Table 1). As discussed below, astrocytes can substantially upregulate TLR expression levels; thus, TLRs in astrocytes may work secondary to PAMP detection by microglia, which are equipped with relatively high levels of TLRs without cell stimulation¹⁶⁷.

TLR3 expression in human astrocytes is induced with various stimuli, including agonists of TLR3 and TLR4 and cytokines, such as TNFα, IL-6, IFNγ, and IL-12¹⁶⁸. In contrast, IL-10, an anti-inflammatory cytokine, does not induce TLR3 expression¹⁶⁸. After TLR3 stimulation, the expression of neuroprotective factors and anti-inflammatory cytokines, such as IL-10, are upregulated, and proinflammatory cytokine expression, such as IL-12p40 and IL-23, is downregulated in a human astrocyte tissue culture setting¹⁶⁸. This results in inhibition of astrocyte growth, promotion of endothelial cell growth, and enhanced neuronal survival¹⁶⁸. In an in vivo brain ischemia mouse model by MCAO, TLR3 agonist poly(I:C) attenuated reactive astrogliosis, reduced brain infarction volume, and improved neurological function¹⁶⁹. These studies suggested that TLR3 in astrocytes could serve as an immunoregulatory receptor. However, another study indicated that TLR3 stimulation with poly(I:C) induced protein expression of TLR2, IL-6, C-C Motif Chemokine Ligand 5 (CCL5), Intercellular Adhesion Molecule 1 (ICAM-1), and Vascular Cell Adhesion Molecule 1 (VCAM-1) by primary mouse astrocytes in tissue culture¹⁷⁰, suggesting TLR3 stimulation enhanced inflammation. The same study also demonstrated that TLR3-stimulated astrocytes slightly triggered antigen-specific CD4⁺ T cell activation ex vivo based on the evaluation of T cell proliferation and IFNγ production¹⁷⁰. A more recent study indicated that human astrocytes treated with IFNγ (although TLR3 agonist was not tested), enhanced the expression of MHC class I and II, as well as co-stimulatory molecules (CD80/86/40)¹⁷¹. These studies suggest the involvement of TLR3 in modulating the CNS environment. However, further study is required to determine the involvement of astrocytic TLR3 in antigen-presentation in vivo as new antifungal host responses.

An early study using primary mouse astrocytes showed expression of Tlr2, 4, 5, and 9 and demonstrated that their gene expression levels could also be induced¹⁷². Various stimuli can induce TLR2 expression on astrocytes. For example, PGN, flagellin, and CpG-DNA highly induce Tlr2 expression in mouse astrocytes ex vivo¹⁷². Bacterial infection with Brucella melitensis also enhanced TLR2 expression in the white matter of rhesus macaques¹⁷³. TLR2 on astrocytes recognizes Streptococcus suis, which causes meningitis, and induces expression of TLR2, TLR6 (but not TLR1), and a downstream proinflammatory response^174,175. Another study showed that TLR2 expression on astrocytes depends on TNFα and NFκB pathways¹⁷⁶. Multiple studies using primary astrocyte tissue culture demonstrated functional outcomes of TLR2 signaling on astrocytes. For example, stimulation of TLR2 with lipoteichoic acid induces the expression of matrix metalloproteinase 9 (MMP-9) in rat astrocytes¹⁷⁷. Ex vivo treatment of mouse primary astrocytes with PGN triggers expression of TNFα, C-X-C Motif Chemokine Ligand 2 (CXCL2), and CCL4, and the involvement of TLR2 in the response was confirmed with Tlr2^−/− astrocytes¹⁷⁶. However, priming may be required for TLR2-mediated functions in astrocytes because TLR2 signaling needed pre-treatment of astrocytes with proinflammatory cytokines, such as TNFα, IFNγ, and IL-1β, for overt TLR2-mediated responses¹⁶⁷. The study used primary mouse astrocytes in tissue culture¹⁶⁷. Sensitization of TLR2 can also be achieved with conditioned media from microglia culture¹⁶⁷, suggesting that astrocytes require crosstalk with microglia or CNS-infiltrated inflammatory cells to efficiently exert outcomes of TLR2 stimulation. For astrocytic TLR2 signaling activation, astrocyte priming may be necessary to induce TLR2 expression to a certain level.

TLR4 on astrocytes appears to be ready to detect ligands because of its constitutive expression. Tlr4 mRNA expression levels without stimulation are already higher than those of Tlr1, Tlr2, and Tlr3 in human astrocytes¹⁶⁸. A study showed that LPS stimulation of primary murine astrocytes induced gene expression of Tnf, Il1b, Il6, Cxcl1, Cxcl2, Ccl2, Ccl7, and Ccl12 as early as 2 hours after stimulation¹⁷⁸, suggesting that TLR4 was present on astrocytes and able to detect LPS¹⁷⁸. Other studies also showed functional TLR4 on astrocytes without pre-treatment. For example, LPS stimulation of murine astrocytes activates NFκB, MAPKs, and JAK/STAT signaling pathways¹⁷⁹, and enhances the expression of inducible Nitric Oxide Synthase (iNOS), IL-6, CCL2, ICAM1, and VCAM1¹⁷⁰. Yet, earlier studies showed conflicting results. One study concluded that astrocytes (and oligodendrocytes) do not express Tlr4 mRNA, despite the positive Tlr4 detection in microglia, using primary cells from newborn rat forebrains without stimulation¹⁸⁰. The study¹⁸⁰ was published in 2002; therefore, the detection of cDNA was performed on an agarose gel, rather than quantitative PCR. Thus, it is possible that detection was not sensitive enough to detect Tlr4 transcripts. In contrast, another study published in 2003 also used agarose gels and clearly showed the Tlr4 cDNA band, although astrocyte stimulation was necessary for visualizing solid Tlr4 cDNA bands¹⁷². A different study showed a very low TLR4 protein expression level in resting mouse astrocytes when detected by flow cytometry¹⁷⁰. Here, it is possible that TLR4 detection by flow cytometry might have been technically challenging if the TLR4 antibody was not sensitive, and the absence of a positive TLR4 staining control prevented evaluation of the TLR4 antibody sensitivity¹⁷⁰. Mouse astrocytes show only mild induction of Tlr4 gene activation with various stimulations, such as LPS, poly(I:C), PGN, flagellin, and CpG-DNA^170,172. The mild induction of Tlr4 is a substantial contrast to that of Tlr2 and Tlr3¹⁷⁰. Tlr4 gene expression in human primary astrocytes is even more resistant to stimulation with inflammatory cytokines, such as IL-1β, TNFα, and IFNγ¹⁶⁸. These basal TLR4 expression levels may be low enough to be missed, and Tlr4 expression is relatively resistant for upregulation. However, it appears that astrocytes can express functional TLR4.

Expression of TLR7 and TLR9 in astrocytes is inducible by ZIKA virus infection¹⁸¹. TLR7 agonist, imiquimod, reactivates astrocyte and induces production of proinflammatory cytokines and chemokines, including IFNβ, TNF, CCL2, and CXCL10¹⁸². Ex vivo TLR9 stimulation of mouse astrocytes with its ligand, CpG-DNA, upregulates the expression of cathelin-related antimicrobial peptide¹⁸³, suggesting that TLR9 signaling is functional in astrocytes. Although ligands and functions for TLR11–13 in astrocytes remain to be addressed, toxoplasmosis in mice upregulates the expression of TLR11 in astrocytes¹⁸⁴.

4.2.2: CLRs in astrocytes

CLR expression by astrocytes appears to be limited. For example, the expression of “Aplec” genes (the “Antigen-Presenting lectin-like receptor gene complex” and including a cluster of C-type lectin (CLEC) receptor gene loci), including Clec4e (encoding Mincle), was not detected in rat astrocytes, while microglia express them when stimulated¹⁸⁵. To the best of our knowledge, no reports showed astrocytic dectin-1 and other major CLRs, involved in fungal detection. Yet, MR is highly expressed in astrocytes in one week-old neonatal mice and decreases its expression level to maintain low levels throughout adulthood¹⁸⁶. Mannose-binding lectin (MBL) is detected in astrocytes of post-mortem brain tissue of patients with HIV encephalitis¹⁸⁷. These studies suggest that astrocytes may not be equipped to directly detect fungal infections through CLRs, although our understanding of CLRs in astrocytes is still not sufficient for this conclusion.

4.2.3: NLRs and ALRs in astrocytes

Astrocytes are known to express some NLRs and AIM2 inflammasome components. For NLRs, NOD1 and NOD2 are expressed in astrocytes¹⁸⁸. NOD2 in retinal astrocytes has been shown to synergize with TLR2 for the recruitment of pathogenic T cells during intraocular inflammation^141,142,189. NLRP3 and NLRC4 inflammasomes are activated in LPC-stimulated mouse primary astrocytes ex vivo, and resulting IL-1β secretion was reported¹⁴⁶, suggesting that astrocytes could behave as inflammatory cells through these inflammasomes. Similarly, ex vivo treatment of mouse primary astrocytes with nigericin, an NLRP3 inflammasome stimulator, also allows IL-1β secretion, although poly(dA:dT)/liposome, an AIM2 inflammasomes stimulator, fails to secrete detectable levels of IL-1β¹⁴⁵. Human primary astrocytes express NLRP2, which can be activated by adenosine triphosphate (ATP) as a DAMP, and process pro-caspase-1 and pro-IL-1β¹⁹⁰. Based on these studies, astrocytes appear to have the capacity to activate various inflammasomes, including NLRs.

Our lab recently reported that astrocytes can also have activated AIM2 inflammasome in the spinal cord during a late stage of EAE¹⁴⁵. Unexpectedly, astrocytes are the main CNS cell type exhibiting AIM2 inflammasome “activation”, and neither microglia nor CNS-infiltrated myeloid cells are sources of any active inflammasomes during EAE¹⁴⁵. Also, the NLRP3 inflammasome is not activated in astrocytes, contrasting to the NLRP3 inflammasome activation and its pathogenic role in EAE in peripheral myeloid cells^{145,191–193}. Interestingly, mouse primary astrocytes with active AIM2 inflammasome release little IL-1β due to globally limited expression of inflammasome components¹⁴⁵. The function of the AIM2 inflammasome in astrocytes is not inflammatory based on limited IL-1β production by astrocytes and more severe EAE in Aim2^−/− mice^145,152,194. Yet, the biological role of active AIM2 inflammasome in astrocytes during EAE is still unknown. Taken together, the roles of activated inflammasomes in astrocytes in vivo are still largely elusive, particularly due to its generally low expression levels of inflammasome components¹⁴⁵. Nevertheless, NLR inflammasomes in astrocytes appear to induce inflammation, while the AIM2 inflammasome in astrocytes may regulate inflammation, or at least its activation is not inflammatory.

4.2.4: RLRs in astrocytes

Reports have shown that RIG-I and MDA5 are expressed in rat, human, and murine astrocytes^195–197. While many studies focus on understanding their roles in CNS injury, few reports address their functionality during infections. Cultured astrocytes in a stretch injury model showed elevated RIG-I and MDA5 expression, leading to the production of IFN-I and an increase in GFAP and vimentin, hallmarks of reactive astrogliosis¹⁹⁵. RIG-I expression in astrocytes is also inducible by hypoxia and infection by vesicular stomatitis virus (VSV)^196,197. In the VSV study, primary human astrocytes produce inflammatory cytokines and neurotoxic mediators¹⁹⁷. In mouse astrocytes, cytosolic poly(I:C) complexed with Lipofectamine activate MDA5, but not RIG-I, and induces the release of IFNβ and IL-6¹⁹⁸.

Astrocytes have functional STING, although the setting in a report did not directly sense cytoplasmic DNA from the astrocyte nucleus¹⁹⁹. Rather, STING was activated by cyclic guanosine monophosphate–adenosine monophosphate (cGAMP), transferred from brain metastatic cancer cells through gap junction channels with connexin-43 (Cx43) through cell-cell contact¹⁹⁹. In addition, the astroglial STING pathway activation elicited the production of IFNα and TNF¹⁹⁹. This intriguing study suggests a possible role of astrocytes as secondary sensors of DNA damage in other CNS-resident cells, such as injured neurons, as well as injured astrocytes, known to be physically connected with healthy astrocytes through Cx43²⁰⁰. The cGAMP-STING signaling in bystander cells is also reported with murine embryonic fibroblasts (MEFs) during virus infection²⁰¹. In this case, volume-regulated anion channels (VRACs) play a role in cell-to-cell transmission of cGAMP to enhance antiviral IFN-I response²⁰¹.

Although astrocytes are not well equipped with various PRRs, compared to microglia with their extensive PRR repertoire, further studies are merited to identify the roles of astrocytes as a primary or secondary cell type to sense infections or even as amplifiers or regulators of infection-triggered cell responses by CNS immune sentinels.

4.3 |. Neurons

Neurons are fundamental cells in the CNS and peripheral nervous system (PNS) responsible for the relay of information throughout the body. Some neuronal PRRs are used to detect infections, but a majority of previous studies focused on the involvement of PRRs in neurons in the light of neuronal death and growth (Table 1; Fig. 2C).

4.3.1: TLRs in neurons

Neuronal TLRs have been studied mainly in the context of neurodegenerative diseases and neurodevelopment. Several studies reported the expression of TLR1–9 in neurons in the CNS and the PNS^{166,202–207}. Interestingly, some neuronal TLRs may function through glycogen synthase kinase 3β (GSK3β), jun-N-terminal kinase (JNK), and PI3K/protein kinase B (AKT) pathways, rather than the canonical NFκB pathway^204,208,209. Expression of TLR2 and TLR4 in neurons is increased in response to IFNγ stimulation and energy deprivation, as well as in response to ischemic brain injury²⁰⁴. The study also reported that TLR2 and TLR4 in neurons enhance proapoptotic signaling, rendering neurons vulnerable to ischemic death²⁰⁴. TLR4 in neurons may have a distinct role or different expression pattern from TLR2 in neurons in AD. For example, only TLR4 expression was increased in neurons when exposed to amyloid β (Aβ) peptide, despite both TLR4 and TLR2 being upregulated in cultured neurons upon exposure to hydoxynonenal (HNE), an AD-related byproduct²¹⁰. TLR4 is also responsible for JNK phosphorylation and caspase-3 cleavage, resulting in neuronal death, upon HNE treatment²¹⁰. This suggests that TLR4 signaling in neurons may be detrimental during AD; but the functional consequence of TLR2 in neurons remains undefined in this setting. Nevertheless, neuronal TLR2 is required for opioid-induced neuronal death via the activation of the GSK3β signaling pathways²⁰⁹, as well as glucose deprivation-induced cell death²⁰⁴. Thus, TLR2 and TLR4 signaling are generally neurotoxic. If fungi stimulate neuronal TLR2 and TLR4, it may be detrimental to neurons rather than eliciting antifungal responses.

TLR3 and TLR5 in primary sensory neurons are involved in controlling itch and pain^211,212. For example, a TLR3 agonist induces action potentials in dorsal root ganglion (DRG) neurons and elicits scratching in wild-type mice but not in TLR3-deficient mice²⁰⁶. TLR5 in DRG neurons was suggested to be a target to treat neuropathic pain through the blockade of Aβ-fibers, which are implicated in mechanical allodynia¹⁰⁶. These studies show that TLR3 and TLR5 regulate neuronal excitability and synaptic transmission.

TLR7 is expressed in neurons, and in vivo neuronal cell death is induced by ssRNA40, an oligoribonucleotide from HIV²⁰⁵. TLR8 in neurons is a negative regulator of neurite growth and an inducer of neuronal apoptosis²¹³. The study also reported that stimulating TLR8 in neurons with R848 (also a TLR7 agonist) downregulates phosphorylated IκB without impacting NFκB p65 activation²¹³. TLR9 in spinal cord neurons is considered involved in the excitotoxic death of neurons, at least partly due to stress in the endoplasmic reticulum²¹⁴. Another study showed that some neurons in mice express TLR11 after encephalitic Toxoplasma gondii infection¹⁸⁴, but the role of TLR11 is unknown.

4.3.2: CLRs in neurons

Spinal cord neurons express low levels of Mincle and induce Mincle expression in response to peripheral nerve injury (PNI). Mincle signaling in neurons appears to induce TNFα production and allodynia induced by PNI^215,216. Neurons also express MR under certain conditions¹⁸⁶, but its functional relevance in neurons remains undetermined. MBL is expressed in human neurons with increased expression during HIV-encephalitis, but the function of MBL is still not clear¹⁸⁷. The roles of CLRs during fungal infections remain unaddressed.

4.3.3: NLRs and ALRs in neurons

NOD1 is highly expressed in neurons of developing mouse brain, and it may function to recognize microbiota-derived PGN that cross the BBB during brain development¹⁸⁸. Rat cortical neurons also highly express NOD1 by an ex vivo oxygen-glucose deprivation and reperfusion treatment²¹⁷. Among this PRR category, neurons have been relatively well studied in the light of inflammasomes. Neurons from naïve mice express detectable levels of NLRP1, caspase-1, and ASC proteins²¹⁸. Human cortical brain neurons also showed the expression of NLRP1, caspase-1, ASC, NLRC4 (IPAF), and AIM2, although NLRP3 expression was not identified²¹⁹. The study demonstrated that ex vivo treatment of human neurons with muramyl dipeptide (MDP), flagellin, and poly(dA:dT) activated caspase-1 in neurons, suggesting functional NLRP1, NLRC4, and AIM2 inflammasomes²¹⁹. Furthermore, IL-1β release from neurons by NLRP1 inflammasome activation induced neuroinflammation, leading to caspase-6-mediated axonal degeneration and Aβ₄₂ overproduction²¹⁹. An earlier ex vivo study showed NLRP1 overexpression in rat primary cerebellar granule neurons led to apoptosis, accompanied by caspase-3 activation, resulting in neuron injury²²⁰. In contrast, a non-inflammasome forming NLR, NLR Family Member X1 (NLRX1), protects neuron-like N2A cells from necrosis²²¹.

AIM2 expression has also been reported in human neuron cultures²¹⁹. The AIM2 inflammasome is activated in rat embryonic cortical neurons, based on the detection of pro-caspase-1 processing, along with the formation of “ASC specks,” which are microscopically detectable signals reflecting the inflammasome polymer complex formation²²². A more recent study indicated the detection of Aim2 transcripts in neurons, along with astrocytes and microglia, during brain development in mice²²³. AIM2 total knockout mice demonstrated increased DNA damage accumulation, which was also identified in neuron-specific caspase-1 knockouts²²³; thus, ASC specks detected in the brain development appear to be activated by AIM2 inflammasome in neurons. Indeed, we identified activated AIM2 inflammasome was also detected in motor neurons expressing Choline Acetyltransferase (ChAT) during a late stage of EAE, although the levels of activated AIM2 inflammasome in neurons are significantly lower than those in astrocytes¹⁴⁵.

4.3.4: RLRs and other PRRs in neurons

Neurons express RLRs in various mammalian models but mainly in the context of viral infections^{208,224–226}. Expression of RIG-I and MDA5 in cultured human and rat primary neurons increases in response to IFN-I; and brain infection by simian immunodeficiency virus also enhances the expression of RIG-I and MDA5 in neurons of macaques^208,224. Furthermore, RIG-I signaling in neurons induces inflammatory molecules, including IL-6, IL-12p70, CCL2, CXCL10, and TNFα, upon Japanese encephalitis virus (JEV) infection²²⁵. A later study showed that RIG-I and STING interact to elicit immunological responses involving IFN-I²²⁶. Together, these reports suggest that neurons are sensitive to detecting viruses through RLRs and elicit antiviral immune responses. However, the roles of neuronal RLRs in the context of fungal infections are still unknown.

4.4 |. Oligodendrocytes and Oligodendrocyte Precursor Cells (OPCs)

Oligodendrocytes are responsible for maintaining and generating the myelin sheath that surrounds axons. They also support neurons by producing neurotrophic factors, which maintain axonal integrity and propagate the action potential along axons. Due to their importance in protecting neurons, oligodendrocytes have become a target of research in neurodegenerative disorders.

To the best of our knowledge, reports on PRR expression in oligodendrocytes and OPCs are limited to a few TLRs (Table 1; Fig. 2D). Among TLRs, oligodendrocytes express TLR2, and the expression is upregulated in MS lesions²²⁷. TLR2 stimulation by its agonists inhibits OPC maturation, while agonists of other TLRs do not²²⁷. Expression of TLR2, and TLR3 to a lesser extent, was also identified in primary rat oligodendrocytes culture²²⁸. TLR3 agonist, poly(I:C), induces apoptosis in oligodendrocytes, while zymosan promotes survival, differentiation, and myelin-like membrane formation in oligodendrocytes²²⁸. Although the study describes zymosan as “a TLR2 agonist”²²⁸, zymosan is a yeast cell wall derivative and stimulates dectin-1 together with TLR2 as well. Thus, dectin-1 signaling may also be involved in the outcomes in zymosan-treated oligodendrocytes, although dectin-1 was not discussed²²⁸. TLR4 was also identified on OPCs and triggers apoptosis of OPCs with Hsp60, a DAMP released by microglia²²⁹, through the TLR4-MyD88-NFκB signaling pathway²²⁹. Another study showed that LPS-mediated TLR4 stimulation is detrimental and toxic for oligodendrocytes in culture¹⁸⁰.

Information on CLRs, NLRs, RLRs, and ALRs in oligodendrocytes and OPCs is limited, except for the study on zymosan-treated oligodendrocytes as mentioned above, although the possible involvement of dectin-1 was not addressed²²⁸. Other studies showed that oligodendrocytes do not express MR¹⁸⁶, but MBL is a CLR expressed by oligodendrocytes of the frontal cortex of HIV-1 infected brains, although the role MBL in oligodendrocytes remains unknown¹⁸⁷.

5 |. RECRUITMENT OF BLOOD-BORNE IMMUNE CELLS TO THE CNS DURING FUNGAL INFECTIONS

Recruitment of immune cells from circulation is essential for fungal clearance in the CNS. Having CNS-infiltrated cells in the CNS adds a broader variety of PRRs and effector functions, including the secretion of cytokines and chemokines, as well as phagocytosis and killing of fungi. However, heightened immunity by CNS-infiltrated cells could cause irreparable damage to the CNS, due to relatively poor regenerative capacity of CNS-resident cells. In addition, as discussed in Section 3, recruited immune cells can contribute to fungal dissemination as vehicles of the Trojan Horse mechanism. Here, we discuss peripheral immune cells recruited to the CNS during fungal infections and their contribution to fungal clearance and CNS pathology.

5.1 |. Neutrophils

Neutrophils serve as the first line of defense against fungal infections and can eliminate pathogens through various mechanisms, including phagocytosis, secretion of antimicrobial factors, generation of oxidative bursts, and release of neutrophil extracellular traps (NETs)²³⁰. In the absence of neutrophils, a condition known as neutropenia, individuals face a higher risk of developing invasive candidiasis²³¹ and aspergillosis²³². In contrast, neutrophil-mediated control of fungal infections is known to cause immunopathology, at least in the peripheral organs^233,234

During disseminated candidiasis, neutrophils are recruited and accumulate in the mouse brain early during infection²³⁵. The number of neutrophils recruited in the brain peaks at 4 days after C. albicans infection, and the kinetics of neutrophil counts in the brain contrasts with a steady increase in the kidney up to day 7²³⁵. Thus, even in the same mouse, the brain appears to limit neutrophil infiltration to a certain point. Recruitment of neutrophils to the CNS during invasive candidiasis relies on Caspase Recruitment Domain Family Member 9 (CARD9) in microglia²³⁶, an adaptor protein downstream of Syk-mediated CLR signaling. CARD9 in neutrophils is also required to produce chemokines that further recruit monocytes and additional neutrophils²³⁶. Neutrophils also play a role to limit the transition from the hyphal form to the yeast form²³⁷.

Cryptococcus infection models using mice^238–240 and zebrafish²⁴¹ showed that neutrophils are the first cells to interact with Cryptococcus in the brain vasculature. Complement C5a receptor signaling is critical for neutrophil recruitment to the brain vasculature, in which neutrophils are the primary cell type to clear C. neoformans²⁴⁰. The clearance of C. neoformans in the brain microvasculature by neutrophils is a crucial step for host protection; thus, increasing neutrophil recruitment to clear arrested cryptococci in the brain microvasculature was proposed as a therapeutic approach to improve cryptococcal meningoencephalitis²⁴⁰. Complement C3 is also critically involved in the recognition and clearance of C. neoformans in the brain²³⁸.

Neutrophil recruitment to the CNS during aspergillosis needs further investigation. Yet, it is reported that neutrophils, recruited to the lung²⁴² and eye²⁴³, mediate the clearance of conidia²⁴⁴ and restrict hyphal growth²⁴³. Based on the studies, it is possible that neutrophils also exert anti-Aspergillus activities in the brain.

5.2 |. Monocytes

Monocytes, which travel through the blood to tissues where they become macrophages or dendritic cells (DCs), are crucial for controlling fungal infections in both mice and humans. (DCs were identified in the cerebrospinal fluid (CSF) of patients with severe cryptococcal meningoencephalitis²⁴⁵, but limited information on brain DCs during CNS mycosis is available.) Monocytes also exert effector functions, such as phagocytosis and secretion of cytokines and chemokines^246,247. Inflammatory monocytes express CCR2, a key chemokine receptor for monocyte migration. CCR2, together with C-X3-C Motif Chemokine Receptor 1 (CX3CR1), is critical for monocyte trafficking along inflamed vessels and subsequent accumulation in the brain²⁴⁸. Although the study is on pulmonary fungal infections, depletion of CCR2⁺ cells in mice reduced host survival and fungal clearance in C. albicans²⁴⁹. The study also suggested that the antifungal activity of CCR2⁺ cells is exerted in the first 48 hours after infection²⁴⁹. In a mouse pulmonary aspergillosis model, CCR2⁺Ly6C⁺ monocytes and their derivatives are required for antifungal CD4⁺ T cell responses in the lung²⁵⁰. More studies describe monocytes in the CNS in C. neoformans infection. Monocytes are recruited and accumulate in the brain vasculature hours after systemic administration of C. neoformans²⁵¹. In this context, monocytes internalize C. neoformans and are taken advantage of as “Trojan Horses” by the fungus to cross the BBB and enter the brain parynchema⁵². Thus, monocytes are generally antifungal, but they are used as vehicles for fungal CNS dissemination.

5.3 |. T cells

The protective role of T cells, particularly CD4⁺ T cells, is well demonstrated in fungal infections. Yet, a limited number of studies described T cells infiltrated in the CNS during fungal brain infections. Among the previous studies, Cryptococcus brain infection is the most well documented among all the fungal infections in the brain, probably due to the urgency to understand prevalent CNS cryptococcosis in immunocompromised individuals, such as HIV/AIDS patients with low CD4⁺ T cell counts²⁵². In comparison to CD4⁺ T cells, CD8⁺ T cells are rather involved in low inocula brain infection by C. neoformans; and CD8⁺ T cells are relatively inefficient in clearing the fungus²⁵³. We also demonstrated that CD8⁺ T cells also do not induce cryptococcal IRIS (C-IRIS), which is induced by CD4⁺ T cells¹⁴. Another ex vivo study showed that CD4⁺ T cells interact with microglia, as possible antigen-presenting cells, better than CD8⁺ T cells²⁵³. These studies suggest a substantially less impact of CD8⁺ T cells in CNS cryptococcosis than CD4⁺ T cells. Nevertheless, CD4⁺ T cell transfer into immunocompromised mice with low degree cryptococcosis did not reduce brain fungal burdens¹⁴, implicating that brain-infiltrated CD4⁺ T cells may also not necessarily be efficient to fight against C. neoformans. However, our study showed that brain-infiltrated CD4⁺ T cells are pathogenic enough to elicit lethal outcomes in C-IRIS, with a majority of T helper cells are Th1 cells¹⁴. Here, IFNγ exacerbates C-IRIS¹⁴. Now, multiple questions remain on brain-infiltrated T cells during fungal infections. For example, which conditions make T cells in the CNS protective or pathogenic during brain fungal infection? Why are CD8⁺ T cells less potent than CD4⁺ T cells in providing either protective or detrimental outcomes in CNS fungal infections? Roles of brain-infiltrated T cells in fungal infections other than cryptococcosis are still largely elusive and merit further investigation.

5.4 |. Other CNS-infiltrated leukocytes

There is limited information on CNS-infiltrated leukocytes other than those mentioned above. Natural killer (NK) cells kill Cryptococcus by a perforin-mediated process²⁵⁴ and show intrathecal expansion and activation during Cryptococcus infection²⁴⁵. Brains from patients with cryptococcosis showed NK cells present within cryptococcomas²⁵⁵. B cells are also recruited to the CNS²⁴⁵ and reduce Cryptococcus dissemination to the brain²⁵⁶. In a mouse model of invasive candidiasis, a study demonstrated a wide range of leukocytes (including NK cells, NKT cells, γδT cells, B cells, CD4⁺ and CD8⁺ T cells, and DCs) detected in the brain²³⁵.

6 |. HOST RESPONSES AGAINST FUNGAL PATHOGENS IN THE CNS

Immune responses in the CNS require tight and masterful regulation to limit collateral damage. Hyperactive inflammatory responses within the CNS are detrimental as they may lead to irreparable tissue damage that can leave long-term and possibly irreversible effects on the host. The CNS has historically been regarded as ‘immune-privileged’, isolated from the peripheral immune system by the BBB. Although a limited number of circulating immune cells are known to patrol the CNS, microglia and perivascular macrophages reside within the CNS²⁵⁷ and are considered the first line of CNS defense against invading pathogens. Additionally, recent advances in neuroimmunology have presented that CNS-resident cells modulate immune responses in the CNS²⁵⁸. However, technical limitations remain in elucidating CNS-resident cells in vivo responses due to the difficulty in reflecting the biological functions of the cells using ex vivo experimental approaches. This section discusses the mechanistic and physiological outcomes of host CNS-resident cells upon detection of fungal pathogens.

6.1 |. CNS immune responses against Candida

Although most Candida infections are seen in mucosal surfaces and on the skin, CNS Candida infections do occur, particularly in the meninges in humans. The bloodstream carries disseminated Candida to all organs, including the CNS. Candida CNS infiltration starts with traversing the BBB²⁵⁹, as described in Section 3.

Once Candida infiltrates the brain, CNS-resident cells respond. Microglia are currently the best-studied CNS-resident cells for their responses and functions. The first article investigating microglia using a mouse model of Candida infection was published nearly 30 years ago. In this study, intracerebral transfer of the BV2 microglial cell line to syngeneic mice inhibited fungal growth in the brain, suggesting the overall protective role of microglia during Candida infection²⁶⁰. A host protective role of microglia could be attributed to their capacity to phagocytose pathogens and serve a sentinel role, by microglia promptly detecting pathogens and producing chemokines and cytokines to recruit inflammatory cells from circulation. Indeed, microglia express CXCL1 upon detection of candidalysin to recruit neutrophils for Candida clearance and release a proinflammatory cytokine IL-1β through NLRP3 inflammasome activation²⁶¹. In microglia, the production of both CXCL1 and IL-1β depends on CARD9²⁶¹. CXCL1 is a critical chemokine for neutrophil recruitment to efficiently clear fungus; and astrocytes assist microglial CXCL1 production by providing additional signals to microglia producing CXCL1 in response to candidalysin²⁶¹. The involvement of astrocytes in the sentinel role of microglia is intriguing because it suggests the active participation of astrocytes at the early stage of antifungal immune responses. At this point, it is unknown if direct contact between microglia and astrocytes is required or if some soluble factors mediate the crosstalk to enhance CXCL1 production by microglia.

As described above, a major role of microglia, at least in Candida infection, is sensing the infection and sending a signal to recruit neutrophils in the brain. The sentinel function of microglia is reminiscent of tissue-resident macrophages in the peripheral organs, such as alveolar macrophages, which secrete chemokines CXCL1 and CXCL2 faster than other cell types during pulmonary cryptococcosis, as our lab previously reported²⁶². We also demonstrated that tissue-resident macrophages produce CXCL1 and CXCL2 more efficiently than blood-borne macrophages for neutrophil recruitment upon C. albicans detection²⁶³. Here, tissue-resident macrophages are poised to induce NFκB activation and secrete CXCL1 and CXCL2 by autophagic digestion of A20 upon detecting Candida²⁶³. Indeed, microglia in naive adult mice also highly express TNF Alpha Induced Protein 3 (TNFAIP3/A20)²⁶⁴. Thus, microglia may also use autophagic A20 degradation as an approach to fulfill their role as CNS immune sentinels.

Candida infection in the CNS also promotes the accumulation of activated microglia, astrocytes, and cleaved Aβ peptide around yeast aggregates and forming fungal-induced glial granulomas¹⁶⁴. Aβ has antimicrobial activities^265,266 and efficiently aids the clearance of Candida by enhancing microglial phagocytosis and antifungal activity¹⁶⁴. However, Aβ is also the key mediator of Alzheimer’s disease pathology²⁶⁷. Therefore, further investigation is warranted to understand the long-term consequence of fungal infections in the brain and their possible connection to the induction of neurodegenerative diseases in later life.

6.2 |. CNS immune responses against Cryptococcus

Dissemination of Cryptococcus into the CNS can result in the development of CM²⁶⁸. Although CM was initially described over a century ago²⁶⁹, immune responses of CNS-resident cells to Cryptococcus remain poorly understood. Cryptococcus forms a large polysaccharide capsule that hides components of the yeast cell wall⁹, resulting in limited and delayed recognition by host immune cells²⁷⁰. Studies using BV2 microglial cells demonstrated that acapsular mutant Cryptococcus yeasts were more susceptible to phagocytosis and clearance than encapsulated yeasts²⁷¹. However, both acapusular mutant and encapsulated yeasts similarly fail to elicit TNFα and IL-1β expression by BV2²⁷¹. Interestingly, the addition of either acapusular or encapsulated yeasts to BV2 cell culture inhibits LPS-mediated TNFα expression²⁷¹. Furthermore, Cryptococcus capsule components, GXM and galactoxylomannan (GalXM), also inhibit LPS-mediated TNFα expression²⁷¹, suggesting that the Cryptococcus capsule can inhibit TLR4 signaling, although it is not clear why acapusular mutant yeasts also inhibited the TLR4 signaling. Here, an antibody against GXM opsonizes C. neoformans, enhancing microglial phagocytosis, limiting Cryptococcus growth²⁷², and inducing chemokine expression of CCL3 (MIP-1α) and CCL4 (MIP-1β)²⁷³. The data highlights the critical involvement of the Cryptococcal capsule in hindering microglia effector functions. At the same time, something other than capsule components appears to be involved in limiting cytokine expression upon detection of C. neoformans. Another study suggested that microglia are, unexpectedly, relatively slow to respond in a mouse model of CM²⁷⁴. Based on the flow cytometry data, seven days post-infection (dpi) was still too early for microglia to upregulate the expression of activation markers, CD11c and major histocompatibility complex class II (MHCII), and the upregulation of the two molecules was found on 14-dpi²⁷⁴. Additionally, the production of cytokines and chemokines, including TNFα, CCL3, and CCL4, required at least 21-dpi, and some molecules took more than 30-dpi for significant upregulation²⁷⁴.

C. neoformans induces immune reconstitution inflammatory syndrome (IRIS), a pathological condition whereby a recovering immune system paradoxically worsens the patient’s condition by excessively responding to a previously acquired C. neoformans infection. IRIS can also be associated with Candida and Aspergillus, but Cryptococcus-associated IRIS (C-IRIS) is most prevalent, and a serious concern, for HIV/AIDS patients. We used a mouse model of C-IRIS, mimicking human C-IRIS observed in immunocompromised patients, and demonstrated that the combination of previous subclinical C. neoformans infection and immune reconstitution by CD4⁺ T cells led to brain swelling by dysregulation of aquaporin-4 expression in astrocytes¹⁴. Microglia activation was also reported in mouse C-IRIS²⁷⁵. In the mouse C-IRIS model, CD4⁺ T, particularly Th1 cells, are involved in the lethal outcomes, but CD8⁺ T cells are not¹⁴. Although it is unknown why CD8⁺ T cells do not elicit the lethal C-IRIS outcomes, another study suggested less involvement of CD8⁺ T cells compared to CD4⁺ T cells in controlling Cryptococcus infection in the presence of microglia²⁵³. The study also indicated a lesser magnitude of MHC-I upregulation than MHC-II upregulation in an IFNγ-stimulated BV2 microglial cell line²⁵³. These studies suggest the participation of IFNγ and CD4⁺ T cells in controlling CNS cryptococcosis, but IFNγ and CD4⁺ T cells could cause collateral damage in the CNS, as observed in C-IRIS^14,275.

Astrocytes also undergo structural and functional changes during CNS mycosis, i.e., astrogliosis. Astrogliosis is associated with destructive parenchymal cryptococcomas in humans²⁶⁸. In patients with CM, astrocytes are the primary source of CXCL10²⁷⁶, one of the most highly expressed chemokines in the CSF²⁴⁵. Human astrocytoma cells have also been demonstrated to engulf cryptocci²⁷⁷. We currently do not fully understand the functions of reactive astrocytes during in vivo Cryptococcus infection. Yet, astrocytes may be involved in limiting CM by nitric oxide production²⁷⁸ and possible CD4⁺ T cell activation by upregulation of human leukocyte antigen (HLA) class II, as demonstrated with astrocytoma cells²⁷⁷. The involvement of neurons and oligodendrocytes in CNS cryptococcosis is not known.

6.3 |. CNS immune responses against Aspergillus

Although Aspergillus infects the brain and forms brain abscesses in immunocompromised individuals²⁷⁹, limited reports are available for host cell responses in CNS aspergillosis. A. flavus infects the microglia cell line CHME-3 and induces expression of various cytokines, Tlr genes, and Mmp genes²⁸⁰. In particular, Il17, Tlr7, Tlr9, and Mmp9 expression is highly upregulated, and the magnitude of gene expression by A. flavus stimulation is generally more significant than that by C. albicans stimulation²⁸⁰. Secreted factors from A. fumigatus, A. flavus, A. terreus, and A. niger. are toxic to primary astrocytes and an astrocyte cell line, but neurons appear to be more sensitive to Aspergillus-secreted toxins²⁸¹. As demonstrated with monocytes and macrophages, conidia and hyphae of Aspergillus potently activate the host complement system to opsonize the fungus and support phagocytosis and killing²⁸². Cerebral aspergillosis induces the expression of complement proteins (C1q, C3, and C4) by neurons, oligodendrocytes, and CNS-infiltrated myeloid cells. However, unexpectedly, microglia only showed a marginal increase in complement protein expression²⁸³. The complement system can enhance various proinflammatory biological functions in the brain, such as cell stimulation and induction of chemotaxis²⁸⁴. At the same time, it can lead to collateral damages in the brain with loss of neurons and oligodendrocytes²⁸⁴.

7 |. CONCLUSIONS

CNS mycoses are emerging as major global health concerns, causing morbidity and mortality, especially in immunocompromised individuals. Although the CNS has been deemed immune-privileged, recent data demonstrate that CNS-resident cell types, particularly glial cells, can detect and respond to invading pathogenic fungi via PRRs and initiate immune responses during infection. Also, immune cell recruitment in the CNS from the circulation during CNS fungal infections was demonstrated. Animal in vivo mycosis models on peripheral organs have provided considerable information, but our knowledge of CNS mycosis is still behind. For example, little is known about what happens to host cells and fungi in the CNS after detecting fungal invasion. Multiple factors contribute to the paucity of our knowledge in CNS mycosis, despite the unacceptable morbidity and mortality rate. The CNS environment is unique: It involves neuro-immune communications and requires tightly regulated immune responses to avoid irreparable damage. In addition, the CNS environment and structure make it technically challenging to assess host cellular responses. New technologies in studying cells in the CNS greatly merit expanding our understanding of CNS mycosis to overcome the challenges.