Microglia efficiently respond to C. neoformans captured in the brain

The brain is the primary target for C. neoformans, which harbors a dense population of resident macrophages known as microglia. To study the responses of microglia during fungal brain invasion, we intranasally infected CX3CR1^gfp reporter mice (microglia labeled) [29] with tdTomato-expressing C. neoformans H99 strain (H99-tdT, serotype A) as a natural infection model. Mouse brains were collected for in situ imaging of cortex region by whole-mount confocal approach to preserve the native interaction information (Fig 1A). C. neoformans forms colonies of varying sizes in the brain, which are surrounded by abundant GFP⁺ cells (Figs 1B and S1A), indicating microglial responsiveness to fungal invasion. The colony periphery exhibited intensive accumulation of GFP⁺ cells, while the inner regions contained substantial GFP⁺ cell debris (Fig 1B). Since monocytes also express GFP in CX3CR1^gfp mice, we used CX3CR1^gfpCCR2^rfp dual-reporter mice to characterize the contribution of monocytes in this model. Although CCR2⁺ monocytes were recruited to the fungal colonies, cells interacting with C. neoformans were predominantly CCR2^-, confirming their microglia identity (S1B Fig). Time course study revealed sustained microglia association as colonies expanded (Fig 1C and 1D).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. Microglia rapidly respond to brain invasion by Cryptococcus neoformans.

(A) The illustration of in situ imaging procedure. (B) Representative images showing the response of CX3CR1⁺ microglia towards C. neoformans. CX3CR1^gfp/+ mice were i.n. infected with 1 × 10⁴ tdTomato-labeled C. neoformans H99 strain (H99-tdT) for 3 weeks. The brains were imaged after tissue clearing. Arrows indicate cell debris from GFP⁺ cells. (C) The association of microglia with C. neoformans in the brain at various time points after i.n. infection. (D) The size distribution of fungal colonies in the brain after i.n. infection at various time points. Each dot represents an individual fungal colony pooled from five mice per group. (E) CX3CR1^gfp/+ mice were i.v. infected with 1 × 10⁷ H99-tdT for indicated time. The brains were sliced into 1 mm coronal sections, and the 3th–7th slice from anterior to posterior were imaged after tissue clearing for quantification (Left). Representative images showing the localization of microglia with C. neoformans after i.v. infection at indicated time points (Right). (F) The percentage of C. neoformans association with microglia at indicated time points. CX3CR1^gfp/+ mice (n = 3 mice/group) were i.v. infected with 1 × 10⁷ H99-tdT for different time points. (G) Higher magnification 3D images showing the association of microglia with C. neoformans in the brain after i.v. infection at indicated time points. (H) The number of fungal cells per colony in the brain following i.v. infection at indicated time points. The data underlying this Figure can be found in S1 Data. Scale bar: 200 μm (B), 50 μm (C and E), 20 μm (G). ** p < 0.01, *** p < 0.001 by one-way ANOVA (D) (F) (H).

https://doi.org/10.1371/journal.pbio.3003642.g001

To overcome the variability of intranasally (i.n.) infection and investigate how early microglia can respond to C. neoformans captured in the brain, we switched to i.v. infection model, in which fungal cells are rapidly trapped in brain capillaries after injection (S1C Fig) [11,28,30]. Remarkably, the majority of C. neoformans cells were already associated with microglia 24 h post-infection (Fig 1E and S1 Movie). Higher-resolution imaging revealed close physical interactions between microglia and the fungi (S2 Movie). Time-course analysis further showed microglial responses begin as early as 6 h post infection (Fig 1F). We further demonstrated that C. neoformans can proliferate in the presence of microglia association (Fig 1G and 1H). Monocyte depletion by Clodronate liposome did not reduce microglial association with C. neoformans (S1D Fig), confirming that the responding microglia are not monocyte-derived. Following recruitment towards C. neoformans, microglia exhibited enhanced GFP expression, making them conspicuous among resting microglia (S1E Fig). These microglia also displayed pronounced morphological changes, represented by increased cell body and retracted dendrites (S1F–S1H Fig), hallmarks of microglia activation [31]. Interestingly, although microglia in the cerebellum typically exhibit longer dendrites than those in the cortex (S1I Fig), they are equally responsive to fungal invasion (S1J Fig).

Notably, high dose fungal infection through the tail vein led to rapid death of the host (S1K Fig). To rule out side effects of high dose i.v. infection of C. neoformans, we also confirmed similar microglial response using a much lower dosage (S2A and S2B Fig). In addition to high-virulent serotype A strain, microglia also respond to serotype D strain with equivalent efficiency (S2C Fig). C. neoformans enlarges its body size within the host environment. Notably, we found the few fungal cells lacking microglial association often appeared smaller, likely due to the lack of viability (S2D Fig). To investigate whether fungal viability influence microglial response, we used heat-killed fungi and confirmed fungal viability is essential for microglial response (S2E Fig). Similarly, inert polystyrene beads of comparable size did not trigger microglia recruitment (S2F Fig). Importantly, microglial recruitment does not require fungal proliferation, as the ras1Δ mutant, defective in growth at 37 °C [32] (S2G Fig), induced microglia recruitment to a similar level as the wild-type strain (S2H Fig). We also noted that, although microglia responded actively towards C. neoformans, they did not show much increase in activation markers like CD11c and MHCii in this acute infection model (S2I Fig). Collectively, C. neoformans captured in the brain elicits rapid microglial responses, which depends on fungal viability but not proliferation.

Microglia respond to C. neoformans prior to its BBB transmigration

Since C. neoformans reaches the brain via circulation, we next investigated the influences to brain vasculature during microglia-fungi interaction. Under steady conditions, microglia are homogenously distributed in the brain parenchyma, with most cells located at a distance from the capillary vessels and a small fraction located juxtavascularly (Fig 2A and S3 Movie). C. neoformans brain infection results in intimate association of microglia with the fungi-residing loci of the capillaries (Fig 2B). Although C. neoformans is well known for its ability to cross the BBB [13,33,34]; surprisingly, we found almost all microglial responses occurred prior to fungal transmigration to the abluminal side of the BBB (Fig 2C). Direct intracerebral injection of C. neoformans into the brain parenchyma triggered limited microglial responses compared to fungi emerging from the vasculature (S3A Fig), emphasizing the importance of the vasculature origin for fungal induced microglial responses. Further labeling revealed that microglial responses were directed almost exclusively toward mother cells rather than daughter cells (Fig 2D). In rare cases where mother cells transmigrated, microglial association with daughter cells was also detected (S3B Fig and S4 Movie).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. Microglia respond to Cryptococcus neoformans prior to its BBB transmigration.

(A) The distribution of microglia in the cortex region of naïve CX3CR1^gfp/+ mouse brain (left) and 3D modeling showing the relative positioning of microglia and blood vessel (right). Microglia located <2 μm from nearest blood vessel were shown in green. The vasculature was labeled by Dylight-649 conjugated Tomato-Lectin. (B) The quantification of shortest distance to capillaries for microglia with or without C. neoformans association. CX3CR1^gfp/+ mice were i.v. infected with 1 × 10⁷ H99-tdT for 24 h. Each dot represents an individual microglial cell pooled from three mice per group. (C) Representative image showing microglia association with C. neoformans prior to fungal BBB transmigration (left). Arrows indicate the fungal-residing loci. The percent location of C. neoformans (inside or outside capillaries) out of all fungi with microglia association was shown to the right. CX3CR1^gfp/+ mice (n = 4 mice/group) were i.v. infected with 1 × 10⁷ H99-tdT for 24 h. (D) Representative images showing microglia tend to associate with mother cells than daughter cells. Red indicates mother cells (arrow). (E) Representative image showing the enwrapping of blood vessels and engulfment of C. neoformans by microglia. 3D surface modeling of C. neoformans, microglia, and vasculature was shown below. (F) Representative image showing the phagocytosis of dye by microglia, suggesting vessel leakage. Since an overdose of lectin was used, unbound free lectin in the circulation may leak outside the capillaries. (G) Representative images showing the leakage of dye from the fungal-residing loci of capillary (upper) and control region without C. neoformans association (lower). Examples of sampling areas (5 μm x 5 μm) were shown as squares. (H) The quantification of dye leakage from capillaries with or without C. neoformans. The data underlying this Figure can be found in S1 Data. Scale bar: 20 μm (A) (C) (D) (G), 5 μm (F), 10 μm (E) (G). *** p < 0.001 by unpaired student t test (B) (C) (H).

https://doi.org/10.1371/journal.pbio.3003642.g002

Microglia actively migrated to and enwrapped the fungi-residing loci of the capillaries, ultimately resulted in fungal phagocytosis (Fig 2E and S5 Movie), likely associated with endothelial cell retraction. The chemokine receptor CX3CR1 has been implicated in the microglia recruitment in other models [31]. However, we found the CX3CR1^gfp/gfp mice, lacking functional CX3CR1, exhibited comparable microglial responses to C. neoformans (S3C and S3D Fig). We next studied the association of C. neoformans with immune cells by flowcytometry. Interestingly, microglia, defined as CD45^intCD11b⁺ (S3E Fig), associated with more fungi than other leukocytes (S3F and S3G Fig). Interestingly, CX3CR1 deficiency did resulted in a modest reduction in fungal association by microglia (S3F and S3G Fig), likely due to reduced phagocytosis.

C. neoformans is known to compromise BBB integrity [13]. Consistently, the capillary-enwrapping microglia often contained inclusion bodies that incorporated the vessel-labeling dye, indicative of local vascular leakage (Fig 2F). After analysis of the leakage around capillary loci with or without C. neoformans, we found that microglia recruitment is strongly associated with vascular leakage (Fig 2G and 2H). Collectively, microglia respond to C. neoformans by migration to and enwrapping the capillaries prior to fungal BBB transmigration, a process associated with capillary leakage.

Interestingly, CX3CR1⁺ monocytes were frequently observed in close proximity to C. neoformans in the capillary (S3H and S3I Fig). However, unlike microglia, monocytes were largely unable to engulf C. neoformans efficiently from the capillary lumen (S3J Fig) and were frequently positioned on one side of the fungi (S3H Fig), consistent with flowcytometry results, possibly due to the complete blockade of capillaries by the fungi. Collectively, microglia respond to C. neoformans prior to its BBB transmigration.

Nonencapsulated fungi fail to elicit robust microglial responses

Intrigued by the robust sensitivity of microglia in detecting C. neoformans, we next questioned whether microglia can detect other nonencapsulated fungal species or not. We found that Saccharomyces cerevisiae failed to induce microglia recruitment in the brain, even after prolonged infection (Fig 3A and 3B). Interestingly, Candida albicans, another common fungal pathogen, grows hyphae in the brain (Fig 3C and 3D). Although hyphae growth of C. albicans can penetrate endothelial layers of the brain (Fig 3E), it did not induce substantial microglia recruitment as C. neoformans (Fig 3F and 3G). Thus, the high efficiency of microglial response is unique to the encapsulated C. neoformans, but not to nonencapsulated fungal species.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Nonencapsulated fungi fail to elicit robust microglial responses.

(A) Representative image showing Saccharomyces cerevisiae cells (red) located inside brain capillary vessel without microglia association. CX3CR1^gfp/+ mice were i.v. infected with 1 × 10⁷ TRITC labeled S. cerevisiae for 24 h. Blood vessels were labeled by Dylight 649-conjugated Tomato Lectin. (B) The percentage of S. cerevisiae with microglia association 24 and 48 h post infection. (C) Representative images showing the growth of hyphae by Candida albicans in the brain at indicated time points post-infection. mice were i.v. infected with 1 × 10⁷ tdTomato-expressing Candida albicans. (D) The length of hyphae growth by C. albicans in the brain at indicated time points post-infection. Each dot represents an individual fungal colony pooled from three mice per group. (E) A representative imaging and 3D reconstructed images viewed from different angles showing the penetration of C. albicans across blood vessel (GFP from Tie2-GFP mice). (F) A representative image showing the lack of association between C. albicans and microglia at indicated time points. CX3CR1^gfp/+ mice were i.v. infected with 1 × 10⁷ tdTomato-expressing C. albicans for 12 and 24 h. The association of microglia and fungi were calculated. (G) Quantification of microglia association with C. albicans. The data underlying this Figure can be found in S1 Data. Scale bar: 20 μm (A) (C) (E) (F). *** p < 0.001 by one-way ANOVA.

https://doi.org/10.1371/journal.pbio.3003642.g003

Microglia association promotes fungal proliferation

To study the consequences of microglia recruitment towards C. neoformans, microglia were depleted using the CSF1R inhibitor PLX3397 mixed in rodent diet. Treatment of PLX3397 diet for 14 days resulted in >90% of microglia depletion (Fig 4A and 4B). Interestingly, despite substantial microglia depletion, the remaining microglia remained responsive to C. neoformans (Fig 4C), although microglial association were less frequent (Fig 4D). Microglia depletion led to slightly smaller fungal colonies (Fig 4E), consistent with a reduction in brain fungal burden, but not in other organs (Fig 4F). Notably, PLX3397 may have profound systemic and cellular effects, especially its monocyte inhibitory roles. To independently validate these findings, we also employed a genetic microglia depletion model using CX3CR1^CreERiDTR mice. Tamoxifen-induced Cre recombinase expression followed by diphtheria toxin (DT) treatment effectively ablated microglia (Fig 4G), and this depletion also significantly reduced brain fungal burden (Fig 4H). These results confirmed that microglia promote C. neoformans growth in the brain [15].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Microglia promote Cryptococcus neoformans proliferation in the brain.

(A) Representative images showing the successful depletion of brain microglial by feeding mice 2 weeks with a diet containing 1,200 mg/kg CSF1R inhibitor PLX3397. (B) Quantification of microglia per field of view (0.2 × 0.2 mm). (C) Representative images showing the association of microglia with C. neoformans with or without treatment of PLX3397 after i.v. infection with 1 × 10⁷ H99-tdT for 48 h. (D) The percentage of C. neoformans with microglia association 48 h post-infection. (E) The number of fungal cells per colony with or without treatment with PLX3397. Each dot represents an individual fungal colony pooled from four mice per group. (F) The fungal burden in different organs after i.v. infection with a low dose (5 × 10⁴) C. neoformans for 3 days with or without PLX3397 treatment. (G) Representative flowcytometry images showing the depletion of microglia (CD45^int CX3CR1⁺) after treating CX3CR1^CreERiDTR or control iDTR mice with diphtheria toxin (DT). Refer to S7C Fig for gating strategy. (H) The fungal burden in different organs after i.v. infection with a low dose 5 × 10⁴ H99 strain for 3 days after treatment with DT. The data underlying this Figure can be found in S1 Data. Scale bar: 50 μm (A), 20 μm (C). * p < 0.05, unpaired student t test (B) (D) (E) (F) (H).

https://doi.org/10.1371/journal.pbio.3003642.g004

Fungal GXM is crucial for microglial response

Capsule production in C. neoformans is governed by several capsule-associated genes, among which, cap59Δ, cap60Δ and cap64Δ are the first identified. We confirmed the lack of capsule for these mutants in vivo (Fig 5A and 5B). Notably, all three mutants showed impaired microglia association (Fig 5C and 5D), with cap59Δ elicited the weakest response followed by cap60Δ, and cap64Δ showed the strongest response among them (Fig 5D and S6–S8 Movies). Heat-killed C. neoformans failed to trigger microglial response, suggesting capsule building blocks, not assembled capsule, mediate the response (S2E Fig). Since cryptococcal capsule consists of GXM and GalXM (Fig 5E), and the cap59Δ mutant is GXM negative but GalXM positive [25]; the minimal microglial response in cap59Δ highlights GXM as the key component. Furthermore, cap59Δ infection did not disrupt the BBB, as indicated by preserved basement membrane (S4A Fig) and intact tight junction protein ZO-1 (S4B Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Microglia association with Cryptococcus neoformans depends on capsule production.

(A) Representative images showing the capsule of C. neoformans wild-type strain and various CAP mutants in vivo. (B) The quantification of capsule thickness in vivo for each strain. Each dot represents an individual fungal cell pooled from three mice per group. (C) Representative images showing the association of microglia with wild type and CAP mutants. CX3CR1^gfp/+ mice were i.v. infected with 1 × 10⁷ TRITC-labeled C. neoformans for various time points. The brains were imaged by 20× objective after tissue clearing. (D) The percentage of C. neoformans with microglia association at 24 and 48 h. CX3CR1^gfp/+ mice (n = 3 mice/group) were i.v. infected with 1 × 10⁷ strains labeled with TRITC for different time points. (E) Schematic depiction of the C. neoformans capsule polysaccharides. Top, the structure of GXM. Bottom, the structure of GalXM. Mannose residues, green spheres; xylose residues, red stars; glucuronic acid residues, half-filled diamonds; galactose, yellow spheres. Single polymer repeat units are shown and all sugars are in the pyranose form unless labeled with f to indicate furanose. The key genes related to the synthesis of GXM and GalXM are highlighted in red. (F) Representative image showing the association of microglia with uge1Δ strain 24 h post i.v. infection, with quantification shown on the right. (G) Representative image showing the association of microglia with uxs1Δ strain 24 h post i.v. infection, with quantification shown on the right. The data underlying this Figure can be found in S1 Data. Scale bars: 20 μm (A) (F) (G), 50 μm (C). * p < 0.05, *** p < 0.001 by one-way ANOVA (D).

https://doi.org/10.1371/journal.pbio.3003642.g005

GalXM, the minor component of capsule, requires UDP-glucose epimerase (encoded by uge1) for production [25]. Accordingly, the uge1Δ strain is GalXM deficient but GXM positive [25]. We found uge1Δ strain triggered comparable microglial responses as wild-type strain (Fig 5F and S9 Movie), indicating GalXM is dispensable for microglia recruitment. Both GXM and GalXM contain xylose residues, whose synthesis requires UDP-xylose encoded by uxs1 gene [35]. Surprisingly, uxs1Δ mutant also induced comparable microglial response as wild-type strain (Fig 5G), suggesting the UDP-xylose production is dispensable for microglia recruitment. Notably, the uge1Δ and uxs1Δ strains exhibited altered capsule thickness in vivo (S5A and S5B Fig), as previously reported [25].

To identify other virulence factors that contribute to microglial response, we performed a screening of virulence factors based on literature for their microglial responses (S1 and S2 Tables). The pka1Δ mutant, encoding the catalytic subunit of cyclic adenosine 5′-monophosphate-dependent protein kinase A, was identified to have dramatically reduced microglia recruitment (S5C and S5D Fig; S10 Movie). This mutant displayed substantially reduced capsule thickness as previously reported (S5E and S5F Fig) [21]. We confirmed that pka1Δ strain produces less GXM (S5E Fig), supporting GXM as the major mediator of microglial response. We next developed a convenient in vivo labeling strategy to quantify the proliferation of non-fluorescent fungi within individual fungal colonies (S6A Fig). Compared to wild type strain which rapidly proliferates (S6B Fig), the acapsular mutants (S6C–S6E Fig), as well as uge1Δ, uxs1Δ and pka1Δ (S6F–S6H Fig), all showed impaired proliferation in the brain. Collectively, these results suggested that GXM production is required for microglial response, and microglia-recruiting capability of C. neoformans does not guarantee in vivo proliferation.

Microglia P2Y12 mediates recruitment towards capillary

Although chemokine receptor CX3CR1 is implicated in microglia recruitment [29,31], its deficiency did not impair microglia movement toward C. neoformans (S3D Fig), suggesting alternative mechanisms are involved. Microglia express multiple purinergic receptors, among which P2X7 and P2Y12 are the most predominant (S7A and S7B Fig). Previous studies have shown that extracellular nucleotides induce chemotaxis of cultured microglia through Gi/o-coupled P2Y receptors [36], and expression of P2Y12 is involved in microglial enwrapping of leaking blood vessel [37]. Since C. neoformans also induces vessel leakage [13], these previous studies inspired us to explore the role of purinergic receptors during microglia recruitment. We found pharmaceutical blockade of P2Y12 purinergic receptor by Clopidogrel significantly inhibited microglia recruitment but not with P2X7 inhibitor oATP (Fig 6A and 6B).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. Microglial P2Y12 mediates recruitment to Cryptococcus neoformans.

(A) Representative images showing the association of microglia with C. neoformans after treatment with the P2Y12 inhibitor Clopidogrel or the P2X7 inhibitor oATP. (B) Quantification of microglia association with C. neoformans 24 h after treatment with purinergic receptor inhibitors. (C) Representative images showing the association of microglia with C. neoformans in CX3CR1^gfp/+ (WT) and CX3CR1^gfp/+ P2Y12^−/− mice. Mice were i.v. infected with 1 × 10⁷ TRITC-labeled C. neoformans for 24 h. (D) The percentage of C. neoformans with microglia association in WT and P2Y12^−/− mice. Mice (n = 4 mice/group) were i.v. infected with 1 × 10⁷ TRITC-labeled C. neoformans for indicated time points. (E) Quantification of fungal proliferation 48 h after i.v. infection of CX3CR1^gfp/+ (WT) and CX3CR1^gfp/+ P2Y12^−/− mice with 1 × 10⁷ H99-GFP (n = 4 mice/group). (F) The quantification of CFU in different organs in CX3CR1^gfp/+ (WT) and CX3CR1^gfp/+ P2Y12^−/− mice after infection with 5 × 10⁴ H99 for 3 days. The data underlying this Figure can be found in S1 Data. Scale bars: 50 μm (A) (C), 20 μm (E). * p < 0.05, * p < 0.05, *** p < 0.001 by one-way ANOVA (B) (D), unpaired student t test (F).

https://doi.org/10.1371/journal.pbio.3003642.g006

To rule out the possible side effects of P2Y12 inhibitors during microglia recruitment to C. neoformans, we obtained P2Y12^−/− mice and bred them with CX3CR1^gfp mice to label microglia. The successful knockout of P2Y12 was confirmed by flowcytometry (S7C Fig). Consistent with drug inhibition results, P2Y12^−/− mice also showed delayed microglial response to C. neoformans (Fig 6C and 6D), which is associated with slightly reduced fungal burden in the brain (Fig 6E). The reduced fungal burden is only seen in the brain but not in other organs (Fig 6F), suggesting the growth-promotive role of microglia [15]. Thus, microglia purinergic receptor P2Y12 is involved in microglia recruitment towards C. neoformans.

Fungal secreted GXM induces endothelial release of nucleotides

As P2Y12 senses both ADP and ATP [38], we next sought to address the source of nucleotides in our model. C. neoformans induces leakage of plasma, however, we found the concentration of ATP in plasma is low, unlikely sufficient to trigger microglia recruitment (S8A Fig). Activated platelets release ADP from dense granules upon blood vessel damage [39]. Thus, we questioned whether platelet is involved in microglial response. However, we did not observe substantial increase of platelets around fungal colonies (S8B and S8C Fig). Moreover, depletion of platelet had no effect in microglial response (S8D and S8E Fig).

ATP is typically sequestered within cells [40,41], and serves as a danger signal upon release [5,6,42]. Although damaged neurons can release ATP, neuron damage unlikely occurs before fungal BBB transmigration. Upon examining the interaction between C. neoformans and its surrounding environment, we found that C. neoformans intimately interacts with endothelial cells (Fig 7A). Although ATP can be released after cell death, endothelial cell death was not detected around C. neoformans even after prolonged time (Fig 7B). Notably, we found GXM released by C. neoformans was uptaken by endothelial cells (Fig 7C), suggesting it may trigger ATP release. To confirm this, we applied purified GXM to brain endothelial cultures for ATP release checking. Although ATP release was marginally increased after GXM treatment, the addition of ecto-ATPase inhibitor ARL [42] enhanced ATP detection, suggesting GXM induced endothelial ATP release is rapidly degraded by surface enzymes (Fig 7D). Additionally, we confirmed the increase in ATP was not attributable to cell death (Fig 7E). Together, microglial response to C. neoformans requires GXM-induced ATP release from endothelial cells, which activates microglial P2Y12 to induce migration.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 7. Cryptococcal GXM stimulates endothelial cells to release ATP.

(A) Vascular endothelial cells in Tie2-GFP mice with and without C. neoformans (TRITC labeled) 24 h after infection. (B) Representative image showing the lack of Sytox Orange (an indicator of cell viability) staining on endothelial cells around C. neoformans (left), artificial injury induced by puncture with a 30G needle was included as positive control (right). (C) Representative immunofluorescence images showing the release of GXM by C. neoformans and its uptake by endothelial cells. (D) ATP release into the supernatant was measured in bEnd.3 endothelial cells treated with cryptococcal GXM (100 μg/mL), with or without 200 μM ecto-ATPase inhibitor ARL 67156 (n = 4). (E) The cell viability was measured for group settings as in (D) by CCK8 assay (n = 4). The data underlying this Figure can be found in S1 Data. Scale bar: 20 μm (A) (B) (C). *** p < 0.001 by one-way ANOVA (D).

https://doi.org/10.1371/journal.pbio.3003642.g007

Ethics statement

All animal procedures were performed in accordance with institutional guidelines approved by the Institutional Animal Care and Use Committee (IACUC) of Shanghai Jiao Tong University (protocol number A2021045) and Shanghai Veterinary Research Institute Chinese Academy of Agricultural Sciences (protocol number SVLAC/XM-M-G23014).

Animals

CX3CR1^gfp/gfp mice (Strain #005582), originally from The Jackson Laboratory (Bar Harbor, ME, USA), were gifts from Prof. Jing Wang (Shanghai Jiao Tong University). CCR2^rfp/rfp mice were gifts from Prof. Xiaoyu Hu (Tsinghua University). P2Y12^−/− mice were gifts from Prof. Junling Liu (Shanghai Jiao Tong University). Tie2-GFP (Strain #003658), CX3CR^CreER/+ (Strain #021160), and ROSA26iDTR (iDTR mice, Strain #007900) mice were purchased from the Jackson Laboratory. Wild-type C57BL/6 mice were purchased from the Beijing Vital River Laboratory Animal Technology Co. The mice were cross-bred to generate CX3CR1^gfp/+ mice, CX3CR1^gfp/+CCR2^rfp/+, CX3CR1^gfp/+P2Y12^−/−, and CX3CR^CreERiDTR mice. For all experiments, male and female mice aged 6–12 weeks (weighing 18–22 g) were used. Animals were housed under a 12 h light/dark cycle at 22 ± 2 °C with ad libitum access to food and water.

Fungal strains

Cryptococcus neoformans strains, including the high virulent serotype A strain C. neoformans var. grubii strain H99 (ATCC 208821) and serotype D strain B3501 (ATCC 34873) were purchased from ATCC and cryopreserved in glycerol stocks at −80 °C. Mutant strains (S1 Table) used in current study were from the C. neoformans var. grubii gene deletion collection purchased by L.W. from the Fungal Genetics Stock Center (FGSC, https://www.fgsc.net) or created in current study. All knockout strains were validated by PCR. In some experiments, the wild-type Saccharomyces cerevisiae S288c strain was a gift from Prof. Xinqing Zhao, and the Candida albicans SC5314 was a gift from Prof. Changbin Chen. For culturing, fungal strains were streaked from the glycerol stocks onto YPD agar plates (Cat#242720, BD) and incubated at 30 °C for 48 h to allow colony formation. Colonies were then inoculated into 15 mL of YPD broth and incubated overnight at 30 °C in an orbital shaker at 200 rpm. Fungal cells were harvested by centrifugation at 500g for 5 min and washed twice with sterile phosphate-buffered saline (PBS). Fungal cell numbers were determined using a hemocytometer. In some experiments, fungi were heat-killed at 55 °C for 30 min. Only freshly harvested fungal cultures grown to log phase were used for all experiments to ensure consistency.

Construction of fluorescent fungal strains

To generate fluorescent C. neoformans for in vivo tracking, a pTET-driven fluorescent protein cassette was integrated into the Safe Haven 2 locus of the fungal genome. The plasmid carrying pTET-driven tdTomato pYZ89 was kindly provided by Prof. Youbao Zhao. The tdTomato coding sequence was replaced with GFP using Gibson assembly. The resulting GFP or tdTomato expression cassettes, along with the hygromycin resistance gene, were fused to the 5′ and 3′ flanking regions (1.0–1.5 kb) of the Safe Haven locus via overlapping PCR. The insertion cassette with fluorescence protein and resistance gene was introduced into the Safe Haven 2 locus of the recipient C. neoformans by the TRACE method. The plasmids encoding Cas9 and sgRNA were from Prof. Linqi Wang. Briefly, C. neoformans cells were cultured to log phase (OD600 = 0.6–1.0), harvested, and washed in ice-cold electroporation buffer (EB: 10 mM Tris-HCl pH 7.5, 1 mM MgCl₂, 270 mM sucrose, 1 mM DTT). After 1 h incubation on ice, cells were resuspended in 250 μL EB. A 45 μL aliquot was mixed with 5 μL DNA (1 μg Cas9, 700 ng sgRNA, 2 ng insertion cassette) and electroporated (2,000 V, 200 Ω, 25 μF, 4.5–5 ms). Cells were recovered in 1 mL YPD at 30 °C for 2 h, pelleted, and plated on selective YPD agar. Colonies were grown at 30 °C for 3 days. The brightest colonies were picked by a 10 μL tip for several generations to the obtain bright fluorescence. All fluorescent transformants were verified by PCR and tested for full virulence in mice.

Chemical labeling of fungi

Chemical labeling is particularly useful for short-term tracking fungal cells, thus was frequently used in our i.v. infection models. For labeling fungi in vitro, the fungal surface was labeled with tetramethylrhodamine isothiocyanate (TRITC). Briefly, 1 × 10⁸ cells were washed twice with PBS and resuspended in 1 mL PBS. TRITC (Sigma-Aldrich, Cat# 87918), prepared as a 10 mg/mL stock in DMSO, was added to a final concentration of 50 μg/mL. Cells were incubated at room temperature for 30 min in the dark with gentle rotation. TRITC labeling also allowed identification of mother cells within proliferating colonies derived from GFP-expressing H99 strains. Following incubation, cells were washed three times with PBS to remove excess dye, resuspended in PBS, and enumerated prior to infection. For in vivo labeling, mice were injected i.v. with 100 μL of 1% Uvitex 2B (Polysciences, Cat# 19517) in PBS 30 min before euthanasia. This method effectively labels C. neoformans with serum contact. All labeling procedures were performed under light-protected conditions to preserve fluorescence signal.

Infection of mice

For the natural infection model, mice were i.n. infected with 1 × 10⁴ fluorescently labeled fungal cells and monitored for 1-, 2-, or 3-weeks post infection. For i.v. infection model, mice were infected with 1 × 10⁷ fungal cells through the tail vein and euthanized at 24 h or 48 h for imaging the fungal interactions with the host. In select experiments, mice were i.v. injected with 5 × 10⁴ for evaluation of CFU 3 days post infection. Colony-forming units (CFUs) were then counted to quantify fungal load. To deliver C. neoformans directly into the brain parenchyma, intracerebral injections (i.c.) were performed using a stereotaxic apparatus (RWD Life Science). Mice were anesthetized with 250 mg/kg Avertin via intraperitoneal injection and secured on the stereotaxic frame. The scalp was disinfected and incised to expose the skull, and bregma was identified as a reference point. A small hole was drilled at defined coordinates (1.0 mm anterior, 1.5 mm lateral, and 2.0 mm deep from bregma for cortical injection). A Hamilton syringe fitted with a 30G needle was used to inject 5 μL of suspension at a rate of 1 μL/min. The needle was held in place for 5 min post-injection to prevent backflow, then slowly withdrawn. The incision was sutured, and mice were given analgesia and monitored on a warming pad until full recovery.

Fungal burden enumeration

To determine organ fungal burden, infected mice at various time points were euthanized to harvest major organs. Each organ was placed into 15 mL tubes containing 2 mL of cold sterile PBS and homogenized using a handheld tissue homogenizer (TH220, Omni International). Serial dilutions of the tissue homogenates were plated onto YPD agar plates, and CFUs were counted after incubation at 30 °C for 36 h.

Whole-mount confocal microscopy (in situ imaging)

To obtain in situ microscopic images of C. neoformans interacting with the host, whole-mount confocal microscopy of fixed brains was performed to minimize disturbance of the native spatial context. For visualizing the brain vasculature, mice were i.v. injected with 50 μL of a 1 mg/mL DyLight 649-conjugated Tomato Lectin (Invitrogen, Cat# L32472) to label endothelial cells. Mice were anesthetized via intraperitoneal injection of 250 mg/kg Avertin working solution. Blood was collected by cardiopuncture, followed by transcardial perfusion with 20 mL of ice-cold PBS containing heparin (10 U/mL) to flush circulating blood cells, followed by 20 mL of 4% paraformaldehyde (PFA) for tissue fixation. Brains were carefully extracted and placed in 1 mL of PBS in a 35 mm glass-bottom imaging dish (#1.5 coverslip thickness; ibidi GmbH), with the dorsal surface facing downward for imaging on an inverted microscope. Images were obtained using Olympus IXplore SpinSR10 spinning disk confocal microscope equipped with a 60× oil immersion objective (NA = 1.30). Visible laser lines at 405, 488, 561, and 633 nm were employed for fluorescence excitation. This confocal system allowed imaging up to a depth of approximately 30 μm. For 3D imaging, z-stacks were acquired at 1 μm intervals over a total depth of 20–30 μm.

Tissue clearing

For imaging larger areas of the brain, the collected brains were cleared by an organic-based tissue-clearing methods based on PEGASOS (polyethylene glycol-associated solvent system) which preserves endogenous fluorescence. Briefly, the method involves steps including fixation, delipidation, dehydration, and refraction index matching in clearing solution. mice were anesthetized and transcardially perfused with 20 ml ice-cold heparin PBS (10 U/mL heparin) followed by 20 mL 4% paraformaldehyde in PBS (pH = 7.4) by a peristaltic pump. Organs were post-fixed in 4% PFA at 4 °C overnight, then washed thoroughly with PBS. Samples were delipidated sequentially in 30%, 50%, and 70% tert-butanol (tB, Sigma-Aldrich, Cat# 360538), each supplemented with 3% Quadrol (Sigma-Aldrich, Cat# 122262) to adjust pH to above 9.5, for 12 h each step. Dehydration was achieved by incubating organs in tB-PEG dehydration solution composed of 70% v/v tB, 27% v/v PEGMMA500 (PEGMMA500, Sigma-Aldrich, Cat# 409529), 3% w/v Quadrol for 48 h at room temperature with gentle rotation. Tissues were subsequently transferred into the BB-PEG clearing solution, composed of benzyl benzoate (BB, Sigma-Aldrich B6630) and PEGMMA500 (7:3 v/v) with 3% Quadrol, for 24 h until fully transparent. Samples were stored and imaged in the clearing solution by Olympus IXplore SpinSR10 spinning disk confocal microscope equipped with Photometrics Prime 95B scientific CMOS camera. Although tissue clearing allows imaging of the entire brain using lower magnification objective; we used a 20× objective for 3D images (step size = 2 μm) to get high-quality images.

Image analysis

The IMARIS 9.8 (Bitplane) software was used for visualization, measurement, 3D modeling, and movie generation. The size or distance was measured at Full Width at Half Maximum from a Gaussian fitted intensity profile. Volume-rendered stack images were created using the “volume rendering” function, optical slices were obtained with the “orthoslicer” tool. For 3D remodeling of structures, the “surface” function was used for each channel. In some analysis, the microglia cells were modeled using the “spot” function. For analysis of the microglia morphology, both the “surface” and “filament” function were used, allowing analysis of microglia volume as well as the total length and number of dendrites. 3D still images were captured via the “snapshot” function, and animations were produced using the “animation” module, movies were exported at a speed of 24 frame/s.

Association of microglia with C. neoformans

Association of microglia with fungi in the brain was analyzed at the level of fungal colonies rather than individual fungal cells. A fungal colony was considered with microglia association if a microglial cell body was in direct contact (<2 μm) with any individual fungal cell within a fungal colony. For each mouse, at least 30 individual fungal colonies were analyzed throughout the brain cortex region or as many as possible for certain mice with fewer events.

Monocyte depletion

For monocyte depletion, mice received i.v. injection of 200 μL Clodronate liposomes (Liposoma B.V.) 24 h before infection. Control mice received an equal volume of PBS liposomes. The efficiency of monocyte depletion was confirmed by flow cytometry of peripheral blood leukocytes.

Immunofluorescence of brain slices

Mice were anesthetized by intraperitoneal injection of 250 mg/kg Avertin (2,2,2-tribromoethanol, Sigma-Aldrich, T48402), followed by sequential transcardial perfusion with PBS and 4% PFA. After post-fixation in 4% PFA overnight at 4 °C, brains were harvested and sequentially dehydrated in 15% and 30% sucrose solutions at 4 °C until the tissue sank. The brains were subsequently embedded in Tissue-Tek OCT Compound (Sakura Finetek, #4583), then snap-frozen in liquid nitrogen. Frozen brains were sectioned at a thickness of 20 μm using a cryostat microtome, and sections were mounted onto Superfrost Plus glass slides (Thermo Fisher Scientific). Slides were fixed in ice-cold acetone for 10 min, followed by blocking with 5% goat serum for 30 min at room temperature. Primary antibody incubation was performed overnight at 4 °C. The following primary antibodies were used: anti-Collagen IV (Chemicon, Cat# AB748), anti-GFP (Invitrogen, Cat# A-11122), anti-GXM (18B7, Sigma-Aldrich, Cat# MABF2069), CD31 (Biolegend, Cat# 102502), and anti-ZO-1 (Invitrogen, Cat# 33-9100). After three washes with PBS containing 0.05% Tween-20, sections were incubated with fluorophore-conjugated secondary antibodies (Invitrogen) for 2 h at room temperature. Following additional washes, nuclei were counterstained with DAPI, and slides were mounted using an antifade mounting medium for imaging by confocal microscope.

Determination of endothelial cell viability in vivo

To determine endothelial cell viability in vivo, Sytox Orange (Invitrogen, Cat# S34861) was administered via i.v. injection at a dosage of 20 μg/mouse 10 min before euthanasia of Tie2-GFP mice previously infected with 1 × 10⁷ C. neoformans. The brain was collected and visualized immediately using confocal microscope.

Microglia depletion

Since microglia survival needs CSF1R signaling, mice were feed with diet containing 1,200 mg/kg PLX3397 (CSF1R inhibitor, Sunwaypharm, Shanghai, China) for 14 days to achieve microglia depletion. For microglia depletion in CX3CR1^CreERiDTR mice. Tamoxifen (Sigma-Aldrich, Cat# T5648) was given to mice via oral gavage as 500 μL of a 20 mg/mL solution in corn oil. Animals received two doses of 10 mg Tamoxifen given 48 h apart. Thirty days later (to allow full replacement of circulatory DTR positive cells), mice were i.p. treated with 1 μg/dose DT (Sigma-Aldrich, Cat# D0564) in 200 μL PBS for three consecutive days. For both treatments, the depletion of microglia was confirmed by fluorescence microscope or flowcytometry.

Flowcytometry

Brains were harvested from euthanized mice following transcardial perfusion with 20 mL of cold PBS. Leukocytes were isolated using a 30% Percoll gradient centrifugation. To block Fc receptors, cells were incubated with anti-CD16/32 antibody (BioLegend, Cat# 101302) on ice for 10 min. Subsequently, cells were stained with fluorescently conjugated antibodies for 30 min, washed with staining buffer, and fixed with 1% paraformaldehyde in PBS. Samples were analyzed using a BD LSRFortessa II flow cytometer. The following antibodies were used: APC-Cy7 anti-CD45 (BioLegend, Cat# 103116), PE-Cy7 anti-CD11b (BioLegend, Cat# 101216), PE anti-CX3CR1 (BioLegend, Cat# 149006), PE anti-P2Y12 (BioLegend, Cat# 848004), BV421 anti-CD11c (BioLegend, Cat# 117343), BV510 anti-MHCii (BioLegend, Cat# 107636), PE anti-CD42d (BioLegend, Cat# 148504). Data analysis was performed using FlowJo software (version 10).

Purinergic receptor blockade

To inhibit purinergic signaling, specific antagonists were used to block P2Y12 and P2X7 receptors. For P2Y12 inhibition, Clopidogrel (MedChemExpress LLC, HY-15283) was administered at a concentration of 50 mg/kg by oral gavage once daily. For P2X7 blockade, oxidized ATP (oATP, Sigma-Aldrich, 505758) was injected intraperitoneally at a dose of 30 mg/kg once daily. Both treatments were administered immediately after C. neoformans infection.

GXM extraction

The GXM extraction methods were modified from previous studies. C. neoformans was cultured overnight in 1 L of YNB medium at 30 °C with shaking. Cells were pelleted, and the culture supernatant containing secreted polysaccharides was collected. Sodium acetate was added to a final concentration of 10% (w/v) with pH adjusted to 7.0 using acetic acid, followed by the addition of 2.5 volumes of ethanol to precipitate crude polysaccharides (GXM and GalXM). The mixture was left overnight, and the resulting precipitate was air-dried and dissolved in 2–3 mL of water. Total polysaccharide content was quantified using the phenol-sulfuric acid method. To purify GXM, the solution was adjusted to 0.2 M NaCl, and a 0.3% CTAB solution was slowly added (3:1 CTAB to polysaccharide, w/w) to selectively precipitate GXM, leaving GalXM in solution. The CTAB-GXM complex was pelleted by centrifugation and washed with 10% ethanol. To remove CTAB, the pellet was dissolved in 1 M NaCl, and GXM was re-precipitated by adding 2 volumes of ethanol. The GXM pellet was then dissolved in 2 M NaCl and dialyzed sequentially against 1 M NaCl overnight and distilled water for one week. The final GXM product was quantified again, lyophilized, and reconstituted in sterile PBS or culture medium for downstream applications.

ATP detection

Mouse brain endothelial cell line bEnd.3 was cultured in 96 well plates until confluency. Cells were then treated with 100 μg/ml of purified capsular GXM for 1 h, with or without the addition of ecto-ATPase inhibitor ARL 67156 (200 μM, Sigma-Aldrich, Cat# A265) for 30 min, the cell culture supernatant was collected for detection of ATP release. ATP detection kit (Sigma-Aldrich, Cat# MAK190) was used for detection following manufacture’s instruction. For the detection of ATP in plasma, mice with or without i.v. treatment with 25 U/mice apyrase (Sigma-Aldrich, Cat# A6535) were euthanized for blood collection 30 min after injection. Platelet-depleted plasma was obtained by centrifugation of freshly collected heparinized mouse blood at 2,000 x g for 15 min.

Cell viability test

Cell viability was assessed using the Cell Counting Kit-8 (CCK-8, Dojindo) following the manufacturer’s instructions. Cells were seeded in 96-well plates at a density of 5,000 cells per well, grown to confluency, and treated as indicated. After treatment, 10 μL of CCK-8 reagent was added to each well containing 100 μL of culture medium. Plates were incubated at 37 °C for 2 h, and absorbance was measured at 450 nm using a microplate reader. Cell viability was calculated after subtracting background absorbance from blank wells.

Statistics

Data are presented as mean ± standard error of the mean. Statistical analyses were performed using GraphPad Prism 10 software (San Diego, CA). For comparisons between two groups, an unpaired two-tailed Student t test was used. For comparisons among more than two groups, one-way analysis of variance (ANOVA) followed by Tukey’s post hoc test was applied to assess group differences. A p-value of <0.05 was considered statistically significant.