VCAM1/VLA4 interaction mediates Ly6Clow monocyte recruitment to the brain in a TNFR signaling dependent manner during fungal infection

Monocytes exist in two major populations, termed Ly6Chi and Ly6Clow monocytes. Compared to Ly6Chi monocytes, less is known about Ly6Clow monocyte recruitment and mechanisms involved in the recruitment of this subset. Furthermore, the role of Ly6Clow monocytes during infections is largely unknown. Here, using intravital microscopy, we demonstrate that Ly6Clow monocytes are predominantly recruited to the brain vasculature following intravenous infection with Cryptococcus neoformans, a fungal pathogen causing meningoencephalitis. The recruitment depends primarily on the interaction of VCAM1 expressed on the brain endothelium with VLA4 expressed on Ly6Clow monocytes. Furthermore, TNFR signaling is essential for the recruitment through enhancing VLA4 expression on Ly6Clow monocytes. Interestingly, the recruited Ly6Clow monocytes internalized C. neoformans and carried the organism while crawling on and adhering to the luminal wall of brain vasculature and migrating to the brain parenchyma. Our study reveals a substantial recruitment of Ly6Clow monocytes to the brain and highlights important properties of this subset during infection.

Introduction provided that mononuclear phagocytes promote brain invasion of the organism [21,26,27]. C. neoformans can survive in monocytes [28,29], and monocytes containing C. neoformans have been found present in the perivascular space of the brain [23,27]. Intravenous administration of C. neoformans-infected macrophages can enhance the brain fungal burden [26], while depletion of monocytes can reduce the brain fungal burden [26,27]. Notably, monocytes harboring C. neoformans have been directly seen to cross a monolayer of brain endothelial cells cultured in vitro [30,31]. However, the dynamic interactions of monocytes with C. neoformans in the brain vasculature in vivo have not been elucidated, mainly because of the technical challenges of the in vivo imaging.
In this study, with the use of intravital microscopy (IVM) we demonstrated that Ly6C low monocytes were predominantly recruited to the brain vasculature using a model of acutely disseminated C. neoformans infection. The influx of Ly6C low monocytes was mainly mediated by the interaction of VCAM1 and VLA4. In addition, TNFR signaling was crucial for Ly6C low monocyte recruitment by enhancing its VLA4 expression. Finally, we showed that Ly6C low monocytes engulfed C. neoformans and carried the organism while crawling along the luminal wall of brain vessels and migrating to the brain parenchyma.

Substantial recruitment of monocytes to the brain after infection with C. neoformans
Under resting conditions, the patrolling behavior of CX3CR1 + monocytes has been visualized in the skin vasculature and mesentery vessels by IVM [7]. Similarly, we observed the crawling behavior of GFP + cells in the brain vasculature of naïve CX3CR1 gfp/+ mice through the craniotomy window using IVM (Fig 1A, S1 Video). The cells could crawl against the blood flow and exhibit a leading edge (front of the cells) (S1 Fig, S2 Video). Interestingly, we observed an impressive increase in GFP + cell recruitment to the brain vasculature after intravenous infection with C. neoformans (Fig 1B, S3 Video). The recruitment of GFP + cells was dramatically elevated 12 h post infection and maintained thereafter (Fig 1C). The increase of GFP + cells in the brain was confirmed by flow cytometry (Fig 1D). We further noticed that GFP + cells were recruited mainly in postcapillary venules but not precapillary arterioles ( Fig 1E). Statistical analysis showed that the number of GFP + cells in postcapillary venules was significantly higher than the number of the cells in precapillary arterioles ( Fig 1E). The recruitment of GFP + cells was almost completely abolished by depletion of monocytes using clodronate liposomes (S2 Fig), suggesting that the recruited GFP + cells were monocytes. Flow cytometry showed that > 95% of recruited GFP + cells expressed CD11b with a small number of cells expressing NK1.1, confirming that the vast majority of recruited GFP + cells were monocytes (Fig 1F &  1G). Thus, C. neoformans infection in the brain induced substantial recruitment of monocytes to the brain.

Ly6C low monocytes are predominantly recruited to the brain
There are two subsets of monocytes in mice [2][3][4]. To characterize the phenotypes of recruited monocytes in the brain, we generated CX3CR1 gfp/+ CCR2 rfp/+ mice and performed IVM on the brain of these mice 24 h following C. neoformans infection. The result indicated that the recruited monocytes expressed various levels of both CX3CR1 and CCR2 (Fig 2A). As it is difficult to clearly differentiate the two subsets based on levels of CX3CR1 and CCR2, we stained monocytes with Alexa Fluor 647 conjugated anti-Ly6C mAb and found that most monocytes expressed relatively low or even negative levels of Ly6C (Fig 2B). The result is consistent with

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection flow cytometry data which showed that the Ly6C low subset comprised the majority of the CX3CR1 + monocyte population recruited to the brain (Fig 2C & 2D). An increase in the Ly6C hi subset was also noticed but less pronounced compared to the Ly6C low population ( Fig  2C & 2D). To further confirm the finding that the majority of recruited monocytes were Ly6C low monocytes, CCR2 rfp/rfp mice (CCR2 was disrupted) were infected with C. neoformans to visualize monocyte recruitment, as CCR2 signaling is required for the Ly6C hi , but not Ly6C low , population to emigrate from the bone marrow [2,6]. As expected, deficiency of CCR2 did not remarkably affect the influx of monocytes to the brain vasculature following C. neoformans infection as visualized by IVM (Fig 2E, S4 Video). In contrast to Ly6C hi monocytes whose recruitment was dramatically reduced in the absence of CCR2 signaling, Ly6C low monocyte recruitment was not affected 24 h post infection and only slightly affected 48 h post infection ( Fig 2F). We speculate that this small difference noted at 48 h post infection may be attributed to the conversion of some Ly6C hi monocytes to Ly6C low monocytes as reported previously [32,33]. In addition to monocyte recruitment, we detected an increase in NK cells ( Fig  2D); however, most of these recruited NK cells did not express CX3CR1 in our experimental setting (S3B Fig). Interestingly, neutrophil recruitment was limited at 24 h post infection ( Fig  2D) which was confirmed by IVM (S4 Fig); this is in sharp contrast to previous studies using LPS treatment [34]. Moreover, we found that the recruitment of Ly6C low monocytes is largely brain specific, as recruitments to other organs (including liver and kidney) are comparable to the systematic increase in the blood (S5 Fig). Finally, we confirmed that some of the recruited Ly6C low monocytes transmigrated to the brain parenchyma (S6 Fig). Taken together, these results demonstrate that Ly6C low monocytes are predominantly recruited to the brain vasculature during C. neoformans infection.

Monocyte recruitment to the brain vasculature is mainly mediated by the VCAM1/VLA4 interaction
To study the molecular mechanisms involved in the recruitment of Ly6C low monocytes to the brain vasculature, we performed IVM on the brain of infected wild type mice and labeled monocytes using Alexa Fluor-647 conjugated anti-CX3CR1 mAb. CX3CR1 + cells in the brain vasculature were stably and brightly labeled by the mAb (Fig 3A, S5 Video). Among various types of adhesion molecules, we first examined whether or not selectins are involved in the recruitment of monocytes during brain infection with C. neoformans. The result showed that blocking E-, P-and L-selectins had a negligible effect on CX3CR1 + monocyte recruitment, suggesting that other molecules mediated the recruitment (Fig 3B). VCAM1 is another molecule capable of mediating rolling and adhesion [35,36]; an increase in VCAM1 expression was detected in the brain after C. neoformans infection (Fig 3C). Importantly, treatment of infected mice with VCAM1 blocking mAb almost abolished monocyte recruitment to the brain vasculature ( Fig 3D). Similarly, a blockade of VLA4 (VCAM1 ligand) also dramatically reduced monocyte recruitment to the brain vasculature ( Fig 3D). S6 Video showed that VLA4 blocking See also S3 Video. (C) The number of GFP + cells recruited to brain postcapillary venules of CX3CR1 gfp/+ mice (n = 5 per time point) at various time points after i.v. infection with 20x10 6 C. neoformans. The number of GFP + cells in each postcapillary venule of 300 μm in length were enumerated. (D) Representative flow cytometry plots of GFP + cells in the brain of naïve and infected CX3CR1 gfp/+ mice 24 h post infection with 20x10 6 C. neoformans. (E) The recruitment of GFP + cells in postcapillary venules comparing to precapillary arterioles in the brain of CX3CR1 gfp/+ mice (n = 5) 24 h post infection with C. neoformans. Left panel: representative IVM images; right panel: quantification. (F) Representative flow cytometry histograms showing CD11b expression on CD45 + GFP + cells (blue solid line) and CD45 + GFPcells (red filled) isolated from the brain of naïve and infected CX3CR1 gfp/+ mice 24 h post infection with 20x10 6 C. neoformans. The percentages of CD11b + cells out of CD45 + GFP + cells were shown in the figure. (G) Representative flow cytometry plots showing NK1.1 expression on a small portion of GFP + cells in naïve and infected mice 24 h after infection. Initially, total CD45 + brain leukocytes were gated. Scale bars: 10 μm. Data are expressed as mean ± SEM and representative of 2 independent experiments. ��� , p<0.001. https://doi.org/10.1371/journal.ppat.1008361.g001

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection antibody efficiently wiped off the attached monocytes from the blood vessel and inhibited their rolling. A blockade of VCAM1 or VLA4 affected not only rolling but also adhesion of monocytes ( Fig 3E). Flow cytometry confirmed that VCAM1 and VLA4 blocking was effective for both Ly6C hi and Ly6C low populations ( Fig 3F). Collectively, these data demonstrate that the interaction of VCAM1 with VLA4 mediates recruitment of Ly6C low and Ly6C hi monocytes to the brain vasculature during C. neoformans infection.

ICAM1 and CD11a but not CD11b are partially involved in monocyte recruitment to the brain vasculature
To study the involvement of β2 integrins in monocyte recruitment, we performed IVM on the brain of C. neoformans-infected wild-type mice, as well as ICAM-1 -/-, CD11a -/-, and CD11b -/mice. A great number of monocytes were observed in the postcapillary venules of infected wild-type and all three types of knockout mice (Fig 4A). A tendency of reduction and a slight but significant drop in the number of recruited monocytes were detected in infected ICAM1 -/and CD11a -/mice respectively; however, monocyte recruitment in infected CD11b -/mice was comparable to infected wild-type mice ( Fig 4B). Of note, a significant reduction was detected in the percentage of adherent monocytes in the brain of infected ICAM1 -/and CD11a -/mice ( Fig 4C). S7 Video demonstrated the behavior of monocytes in CD11a -/mice; most monocytes underwent fast rolling with very few cells adhering to the vessel, suggesting a role for CD11a in monocyte adhesion. Flow cytometry confirmed that there was significantly fewer Ly6C low monocytes recruited to the brain in ICAM1 -/and CD11a -/mice 48 h post infection compared to wild-type mice ( Fig 4D).
TNFR signaling is required for monocyte recruitment to the brain vasculature via enhancing VLA4 expression on monocytes TNFR signaling plays an important role in the recruitment of leukocytes [34, 37, 38]. Increase in TNF-α mRNA expression was detected in the brain of C. neoformans-infected mice ( Fig  5A). Infected TNFR -/mice exhibited dramatically reduced monocyte recruitment to the brain vasculature ( Fig 5B, S8 Video). Flow cytometry confirmed that deficiency of TNFR led to a significant drop in the numbers of Ly6C hi and Ly6C low monocytes in the brain of infected mice ( Fig 5C). Thus, TNFR signaling was critically involved in monocyte recruitment to the brain vasculature. Next, we examined how TNFR signaling affects monocyte recruitment during brain infection with C. neoformans. The results indicated that infection with C. neoformans led to an increase of VCAM1 expression on endothelial cells (Fig 5D) in both wild-type and TNFR -/mice ( Fig 5E). In contrast, C. neoformans infection induced a dramatic elevation of VLA4 expression on monocytes in wild-type mice but only a slight increase in TNFR -/mice, resulting in a significantly higher expression of VLA4 on monocytes of infected wild-type mice compared to TNFR -/mice ( Fig 5F). This result suggested that TNFR signaling on Ly6C on recruited monocytes (arrow) in brain postcapillary venules of CX3CR1 gfp/+ CCR2 rfp/+ mice 24 h after infection with 20x10 6 C. neoformans. For labeling of Ly6C, mice were i.v. injected with 2 μg AF647-anti-Ly6C mAb through the tail vein 5 min before imaging. (C) Representative flow cytometry plots of CX3CR1 + Ly6C hi and CX3CR1 + Ly6C low monocyte subsets in the brain of naïve and infected C57BL/6 mice. Gated on CD45 + cells. The expression of CD11b on CX3CR1 -(filled) and CX3CR1 + (solid) populations was shown on the right. C57BL/6 mice were i.v. infected with 20x10 6 C. neoformans and 24 h later brain leukocytes were isolated and analyzed by flow cytometry. (D) Flow cytometry analysis of the numbers of different subsets of cells recruited to the brain of C57BL/6 mice (n = 4 per group) 24 h after infection. Ly6C hi monocytes were defined as CD45 + Ly6G -NK1.1 -CD11b + CX3CR1 + Ly6C hi , Ly6C low monocytes as CD45 + Ly6G -NK1.1 -CD11b + CX3CR1 + Ly6C low , NK cells as CD45 + NK1.1 + , neutrophils as CD45 + CD11b + Ly6G + .

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection monocytes regulated the recruitment of the cells. To confirm this result, we purified bone marrow monocytes from wild-type and TNFR -/mice (both CD45.2 background) using negative selection method and labeled the cells with lipophilic dye CellVue and PKH26 respectively. The cells were mixed at 1:1 ratio and adoptively transferred into congenic CD45.1 recipient mice. The recipient mice were then i.v. infected with 20x10 6 C. neoformans and euthanized 24 later to analyze brain monocyte recruitment by flow cytometry. The result showed that TNFR -/monocytes were recruited less compared to wild-type monocytes when equal numbers of monocytes were transferred during C. neoformans infection ( Fig 5G). Collectively, TNFR signaling is critically involved in monocyte recruitment to the brain vasculature through enhancing VLA4 expression on monocytes during C. neoformans infection.

Ly6C low monocytes engulf C. neoformans and carry the organism while crawling on and adhering to the vessel wall and migrating to the brain parenchyma
Having addressed the mechanism underlying monocyte recruitment to the brain vasculature, we next directly visualized the dynamic interactions of recruited monocytes with C. neoformans and the vessel wall in the brain vasculature of CX3CR1 gfp/+ mice in real-time using IVM. The results showed that C. neoformans was engulfed by CX3CR1 + monocytes and monocytes carrying C. neoformans were frequently seen to crawl on postcapillary venules ( Fig 6A, S7A  Fig, S9 Video). In addition, CX3CR1 + monocytes carrying C. neoformans were observed to firmly adhere to the luminal side of postcapillary venules (Fig 6B, S10 Video). We also detected CX3CR1 + monocytes carrying C. neoformans which were in the process of crossing the vessel wall or were located in the brain parenchyma close to postcapillary venules (Fig 6C), supporting a role of monocytes in the transport of C. neoformans from the circulation to the brain parenchyma. Next, we examined the interactions of monocytes with C. neoformans in brain capillaries. In contrast to postcapillary venules, we found fewer CX3CR1 + monocytes in capillaries; the monocytes crawled slowly in capillaries (S11 Video) and adopted a rod-shaped morphology when passing through (Fig 6D, S11 Video). As shown in Fig 6D, monocytes in close proximity to, attaching to and engulfing C. neoformans were seen in capillaries. In addition to a single organism inside monocytes (Fig 6D), we also observed multiple yeasts within one monocyte (S7B Fig). As shown in S7B Fig, one of the three yeasts within the monocyte appeared to spread from the monocyte to an endothelial cell, a phenomenon of cryptococcal cell-to-cell spread described in vitro [28,29]. Finally, we examined the phenotypes of monocytes carrying C. neoformans. Interestingly, only Ly6C low monocytes were detected to carry C.

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection neoformans, although Ly6C hi monocytes were also present in the postcapillary venules ( Fig  6E). Flow cytometry confirmed that monocytes carrying C. neoformans were almost exclusively Ly6C low subset in the brain of infected CX3CR1 gfp/+ mice (Fig 6F, S7C Fig). It is important to point out that that the majority of yeast cells in the brain were outside phagocytes (S7C Fig). Taken together, these results demonstrate that Ly6C low monocytes internalize C. neoformans and bring the organism to the brain parenchyma.

Discussion
Monocytes play a prominent role in infections and inflammation [1,2]. The dynamics and molecular mechanism of monocyte recruitment and the role of monocytes, particularly Ly6C low monocytes, during brain fungal infections remain largely unknown. In this study, with the use of IVM, we observed a remarkable influx of monocytes in the brain vasculature starting at 12 h post infection with C. neoformans. Interestingly, the Ly6C low subset comprised the majority of the monocytes recruited to the brain, although an influx of Ly6C hi monocytes was also noted. In this regard, recent data demonstrated that Ly6C low monocytes were remarkably recruited to the joint during the development of arthritis in mice and differentiated into inflammatory macrophages, promoting disease pathogenesis [13]. It was reported that Ly6C low monocytes recruited neutrophils in mesenteric veins, driving vascular inflammation [12]. In addition, Ly6C low monocytes were seen to interact with neutrophils through cell-cell contact in the vasculature of kidneys, promoting neutrophils activation and ROS generation [8,11]. In contrast, the influx of neutrophils to the brain was limited at 24 h and 48 h post infection compared to monocytes in our experimental setting. Neutrophils are usually the first cells that are rapidly recruited to infection sites; however, the frequency of neutrophils was low in the brain vasculature even at earlier time points following brain infection with C. neoformans [39].
It is important to note that Ly6C hi monocytes have been implicated to play important roles during pulmonary infection of C. neoformans [40,41]. CCR2/CCL2 axis is involved in human brain infections with C. neoformans [42,43]. The relatively less recruitment of Ly6C hi monocytes to the brain observed in our experimental setting might be attributed to the delay of CCR2 + Ly6C + monocyte mobilization from the bone marrow. Indeed, accumulation of CD11b + Ly6C + myeloid cells was observed in the brain of infected mice only starting on 14 days after intravenous infections of C. neoformans strain 52D [44]. In this regard, we found that an initial pulmonary infection followed by intravenous infection two weeks later significantly enhanced Ly6C hi monocyte recruitment to the brain (S8 Fig).
Leukocyte migration depends on adhesion molecules [2]. Identifying the adhesion molecules involved in Ly6C low monocyte recruitment to the brain would be of interest for therapeutic purposes in order to selectively block the trafficking of the cells. The adhesion molecules in monocyte trafficking have been extensively studied using in vitro systems [2,[15][16][17]45]. However, relatively less is known about the adhesion molecules for monocyte recruitment in vivo. During Leishmania major infection, monocyte migration through inflamed dermal venules relies on interactions of PSGL-1 with P-and E-selectins, and of L-selectin with PNAd, while migration through lymph node high endothelial venules is dependent on L-selectin-PNAd interactions [46]. In the setting of atherosclerosis, P-selectin, E-selectin, and VCAM1 are

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection involved in monocyte rolling on inflamed aortic endothelium [47]. Blocking CD11b or ICAM1 using antibodies inhibits Ly6C hi monocyte recruitment to the liver during infection with Listeria monocytogenes [48]. P-selectin has been shown to mediate Ly6C hi monocyte recruitment to the brain in a CCR2 dependent manner during peripheral organ inflammation [38,49]. In the current study, we reveal that Ly6C low monocyte recruitment to the brain was mainly mediated by the interaction of VCAM1 expressed on endothelial cells and VLA4 expressed on the monocytes following infection with C. neoformans. This is in contrast to steady condition in which crawling of Ly6C low monocytes along endothelial cells of dermis is mediated by CD11a [7]. These results suggest that the adhesion mechanisms used by monocytes may differ, depending on the subset of monocytes, the type of tissues, the infectious agent, and the inflammatory state.
In this study, we have shown that TNFR signaling played an essential role in monocyte recruitment to the brain via enhancing VLA4 expression on monocytes. Previous studies have shown that intraperitoneal administration of TNF-α induced marked neutrophil rolling and adhesion in the brain microvasculature via activating brain endothelial cells and enhancing their expression of P-and E-selectin [37]. Accordingly, TNFR -/mice displayed significant reductions in rolling and adhesion of neutrophils in the brain in response to LPS [34]. These results demonstrated that TNFR signaling is critical for the neutrophil recruitment to the brain via acting on the brain endothelium. In contrast, in our experimental setting of brain infection with C. neoformans TNFR signaling mediated monocyte recruitment through acting on monocytes. Thus, although TNFR signaling is a central mediator of both neutrophil and monocyte recruitments, the underlying mechanisms appear distinct, reflecting the difference in adhesion molecules used by these cells for recruitment to the brain. Following C. neoformans infection, we detected enhanced expression of TNF-α in the brain. Microglia is a potential producer of TNF-α during brain inflammation [50]. In addition, it has been recently shown that Ly6C low monocytes rapidly secrete a large amount of TNF-α during inflammation [11,12] and infection with L. monocytogenes [7]. The cellular source of TNF-α in the brain infected with C. neoformans deserves further investigation.
C. neoformans can enter the brain via direct invasion of the brain endothelium [22][23][24][25]. There is also evidence that monocytes can act as Trojan horse, contributing to brain invasion by C. neoformans [23,26,27]. Notably, live cell imaging has been recently performed to directly visualize the interactions of monocytes containing C. neoformans with a monolayer of human brain endothelial cells cultured in vitro, leading to transmigration of the infected monocytes [30, 31]. Taking advantage of IVM, we have visualized the dynamic interactions among recruited monocytes, C. neoformans and brain vasculature in live animals. Following phagocytosis of C. neoformans, monocytes carrying the organism crawled on and adhered to

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection postcapillary venules and were occasionally seen to migrate to the brain parenchyma. Interestingly, the monocytes carrying C. neoformans were almost exclusively Ly6C low monocytes. Given that both Ly6C hi and Ly6C low monocytes have the potential to internalize C. neoformans, we cannot exclude the possibility of transition of Ly6C hi monocytes into Ly6C low monocytes following phagocytosis of the organism, as in situ monocyte transition (from Ly6C hi to Ly6C low ) has been previously documented [32,33]. It is worthy to note that Ly6C low monocytes have been recently shown to carry ingested amyloid beta as they crawl on the luminal wall of brain vasculature [51]. In addition to postcapillary venules, we also visualized the behavior of monocytes in capillaries. Monocytes adopted a rod-shaped morphology when crawling within capillaries. Monocytes carrying C. neoformans were also observed in capillaries. It has been reported that rod-shaped Ly6C low monocytes patrol the brain capillaries and give rise to perivascular macrophages under inflammatory conditions [52]. It remains to be determined whether monocytes carrying C. neoformans in capillaries transmigrate to the brain parenchyma.
The transcription factor Nr4a1 controls Ly6C low monocytes differentiation and survival [53][54][55]. Consequently, Nr4a1 -/mice displayed dramatically diminished Ly6C low monocyte recruitment to the brain after fungal infection (S9A Fig). We hypothesized that the reduced Ly6C low monocyte recruitment in Nr4a1 -/mice led to a decreased fungal burden in the brain. We repeatedly observed a slight reduction of brain CFU (~15%) in infected Nr4a1 -/mice compared to infected wild-type mice (S9B Fig). Blocking VCAM1 or VLA4 also slightly reduced brain fungal burden (S9C Fig). However, the reduction of brain CFU did not reach statistical difference. ICAM1 deficiency enhanced brain CFU (S9D Fig), which may reflect less killing of C. neoformans in the brain vasculature due to reduced recruitment of neutrophils [56]. It is noteworthy that, in addition to hijacking of phagocytes, C. neoformans can use other mechanisms including transcytosis and paracytosis to invade the brain [57]. It is likely that Ly6C low monocytes make less contributions to C. neoformans invasion into the brain in our model system. Although Ly6C low monocytes were seen to carry C. neoformans while crawling on and adhering to the luminal wall of brain vasculature and migrating to the brain parenchyma, further investigation is required to determine the relative contribution of this subset compared to other mechanisms involved in brain invasion by C. neoformans.
In summary, there was a substantial recruitment of monocytes in the brain following infection with C. neoformans. Interestingly, the recruitment was primarily of Ly6C low monocytes and mainly mediated by the interaction of VCAM1 with VLA4 in our model system. TNFR signaling played an essential role in monocyte recruitment by enhancing VLA4 expression on monocytes. Although the conclusions were mainly made using high dose infection, we proved that the mechanisms also apply to low dose infection (S10 Fig). Finally, in addition to the previously described anti-and pro-inflammatory properties, we showed that Ly6C low monocytes neoformans. See also S10 Video. (B) A representative IVM image showing that a GFP + monocyte (green) carrying C. neoformans (red) was attached to the luminal side of a postcapillary venule in the brain of infected CX3CR1 gfp/+ mice. IVM was performed 18 h post infection with 20x10 6 C. neoformans and the blood vessel (blue) was labeled with AF647-conjugated BSA. See also S11 Video. (C) Representative IVM images showing that GFP + monocytes (green) carrying C. neoformans (red) were in the process of crossing the vessel wall (left panel) or were located outside the postcapillary venule (right panel). IVM was performed on the brain in CX3CR1 gfp/+ mice 18 h post infection with C. neoformans and the blood vessel (yellow) was labeled with AF647-conjugated BSA. (D) Monocytes (green, arrowhead) were seen in close proximity to C. neoformans (red, arrow) in capillary vessels (blue, upper panel), or attaching to C. neoformans (middle panel), or engulfing C. neoformans within capillary vessels (lower panel). Immunohistochemistry was performed for the brain tissues of C57BL/6 mice 18 h after infection with 20x10 6 C. neoformans. (E) Representative IVM images showing GFP + monocytes (green) carrying C. neoformans (red, arrow) were Ly6C (blue, AF647-anti-Ly6C mAb) negative. IVM was performed on the brain of CX3CR1 gfp/+ mice 18 h after infection with C. neoformans. (F) Representative flow cytometry histograms showing the expression level of Ly6C on monocytes carrying C. neoformans. Brain leukocytes were purified from CX3CR1 gfp/+ mice (n = 5) 18 h after infection with 20x10 6 Uvitex 2B labeled C. neoformans. Monocytes carrying C. neoformans were defined as CD45 + Uvitex 2B + GFP + cells (upper panel). The expression of Ly6C on monocytes carrying C. neoformans (CD45 + Uvitex 2B + GFP + ) was compared to Ly6C expression by Ly6C hi monocytes, neutrophils, and Ly6C low monocytes (lower panel). Scale bars: 10 μm. https://doi.org/10.1371/journal.ppat.1008361.g006

PLOS PATHOGENS
Ly6C low monocyte recruitment to the brain during fungal infection were capable of internalizing C. neoformans and carrying the ingested pathogen as they crawled along the luminal wall of the brain vasculature and migrated to the brain parenchyma.

Ethics statement
This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals from National Research Council. Animals study protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of University of Maryland, College Park under identification number R-15-17 and R-SEPT-18-51.

C. neoformans and infection
The encapsulated C. neoformans strain H99 (serotype A) was obtained from the ATCC (Cata-log# 208821). The GFP-expressing H99 strain was provided by Dr. Robin May (University of Birmingham). The tdTomato-expressing H99 strain was generated and provided by Dr. Xiaorong Lin (University of Georgia). In some experiments, the yeast cells were labeled by tetramethylrhodamine [25] or Uvitex 2B [59] before use. The organisms were grown to log phase in Sabouraud's dextrose broth at 32˚C with gentle rotation overnight and washed twice in sterile PBS. For infection, mice were infected with 20x10 6 C. neoformans via the tail vein.

Intravital microscopy
Intravital microscopy (IVM) was performed on the brain vasculature as previously described [25,34]. Briefly, mice were anesthetized by i.p. injection of a mixture of 10 mg/kg xylazine and 200 mg/kg ketamine hydrochloride in 200 μl PBS. After confirming anesthesia, the skin over the skull was moistened with oil and removed using surgical scissors. A surgical craniotomy was then performed using a high-speed drill (Ideal Micro-Drill) with a tip diameter of 0.9 mm to generate a circle (5 μm in diameter) on the right side of the skull. The cranial window was opened by gently removing the covering skull to expose the underlying pial vasculature. After the surgery, the mouse was fixed on a customized stage for IVM. A drop of filter-sterilized artificial cerebrospinal fluid (119 mM NaCl, 26.2 mM NaHCO 3 , 2.5 mM KCl, 1 mM NaH 2 PO 4 , 1.3 mM MgCl 2 , 10 mM glucose, 2.5mM CaCl 2 ) was applied to the exposed portion of the brain and maintained throughout the experiment. The body temperature of the animals was maintained at 37˚C throughout the experiment using a heating pad with a sensor probe inserted in the murine anus (Harvard apparatus). The FITC channel which detects GFP signal or tissue auto-fluorescence was first used to locate the field of view (FOV) of interest. Once the FOV was located, a video was captured through a 40x water immersion lens at a rate of 1 frame/ PLOS PATHOGENS Ly6C low monocyte recruitment to the brain during fungal infection second for later analysis. A typical FOV of interest included a postcapillary venule of 30-70 μm in width and at least 300 μm in length as described previously [34]. In some experiments, mice were injected with 2 μg Alexa Fluor-647 conjugated anti-CX3CR1 or anti-Ly6C mAbs via the tail vein to label monocytes 5 min before visualization. To block the adhesion molecules, mice were injected with 100 μg blocking mAbs through the tail vein at indicated time.

Video analysis
The numbers of rolling and adherent (including crawling) monocytes were quantified from IVM video analysis using the Zen software. The definition of rolling and adhesion followed previous publications [36, 60,61]. Briefly, rolling monocytes were defined as those monocytes moving at a velocity slower than erythrocytes within a given vessel and showing discreet interaction with the vascular wall which interrupted their circulatory movements for less than 30 seconds. Adherent monocytes were defined as those monocytes continuously interacting with the blood vessel for more than 30 seconds. At a given time point, the total number as well as rolling/adherent monocytes (either GFP + or CX3CR1 + ) within a blood vessel of 30-70 μm in width and 300 μm in length were enumerated.

Brain leukocyte preparation
After anesthesia of the mice, blood was taken by cardiopuncture from the right ventricle of the mouse heart. Then, the inferior vena cava was cut open by surgical scissors, and 20 ml cold PBS was used to perfuse the body from the left ventricle to remove free blood cells in the circulation. After perfusion, the brain (including the cortex and cerebellum) was harvested into 15 ml tubes containing 2 ml RPMI 1640 medium, minced into small pieces with a 1 ml pipet tip. Collagenase IV (Worthington) was added to a final concentration of 1 mg/mL and the tissues were incubated for 30 min at 37˚C with gentle shaking. After enzymatic digestion, 100% Percoll solution was added to the tissue solution to a final concentration of 30% Percoll (GE), and filtered through a 70 μm cell strainer on ice. The tissues in 30% Percoll were layered onto a new 15 ml tube containing 2 ml 80% Percoll and centrifuged at 1500g for 15 min without brake. The leukocytes on the 80% and 30% interface were collected into a new 15 ml tube, washed with PBS; treated with ACK buffer for red blood cell lysis, and resuspended in 200 μl cold flow staining buffer (1% BSA in PBS with 0.05% sodium azide). Cells were counted under hemocytometer and ready for staining.

Flow cytometry
For flow cytometry, up to 1x10 6 brain leukocytes were suspended in 50 μl flow cytometry staining buffer. Fc receptors were blocked using anti-CD16/32 mAb (93; eBioscience) at 10 μg/ ml for 20 min on ice. The cells were then stained by fluorophore conjugated mAbs for 30 min on ice. After washing, the cells were fixed in 1% PFA for 20 min; washed and resuspended in 200 μl flow staining buffer for detection using FACSCanto II flowcytometer (BD Biosciences). A minimum of 100000 events were recorded and the data obtained were analyzed using

Immunohistochemistry
After infection, mice were euthanized, and the brain was removed after perfusion as described above. The brain was immediately frozen in OCT compound and cut on a cryostat microtome at a thickness of 5-μm sections onto coated glass slides. Tissue sections were fixed in ice cold acetone for 10 minutes. Sections were then incubated with 3% goat serum in PBS, followed by incubation with rabbit-anti-mouse collagen IV (Invitrogen), rat-anti-mouse F4/80 (eBioscience), and a mouse mAb specific for cryptococcal polysaccharide (E1, a gift from Françoise Dromer, Institut Pasteur, Paris) at 4˚C overnight in a humidified chamber. After 3 washes, sections were incubated for 30 minutes with AF647 goat anti-rabbit IgG (Invitrogen) to delineate brain micro-vasculature, and with AF488 goat anti-rat IgG (Invitrogen) to identify monocytes, and AF555 goat anti-mouse IgG (Invitrogen) to stain the yeast cells in the brain microvasculature. The sections were rinsed and mounted with fluorescence anti-fading medium (KPL).

Monocyte adoptive transfer
Monocytes used for adoptive transfer were isolated from the bone marrow of WT and TNFR -/mice (both CD45.2 background) using monocyte isolation kit from Miltenyi Biotec (Catalog# 130-100-629) following manufacturer's instructions. The isolated monocytes from WT and TNFR -/mice were separately labeled with lipophilic membrane dye CellVue Claret Far Red and PKH26 (both from Sigma-Aldrich) and mixed at 1:1 ratio. 2x10 6 mixed monocytes were transferred to each congenic mouse with CD45.1 background 2 h before infection with 20x10 6 C. neoformans; mice were euthanized 24 h post infection to examine the recruitment of donor monocytes.

Quantitative PCR
A small piece of the brain was collected for RNA extraction by Trizol (Invitrogen) according to the manufacturer's instructions. The cDNA was synthesized by SuperScript IV First-Strand Synthesis System (Invitrogen). qPCR was performed on a Bio-Rad CFX96 qPCR instrument using SYBR Green PCR Master Mix (Applied Biosystems). The PCR primers for TNF and VCAM1 are as follows: VCAM1-F (AGT TGG GGA TTC GGT TGT TCT)

Statistics
Data were expressed as mean ± SEM. Evaluation of statistical significance was performed using Graph-Pad Prism 5 software (San Diego, CA). If only two groups were involved in the analysis, an unpaired two-sided Student's t-test was performed; for comparisons of more than two groups, one-way analysis of variance (ANOVA) was performed followed by Tukey's posthoc test to determine significance among groups. In both cases, p <0.05 is considered significant. (AVI) S11 Video. IVM video showing that a monocyte (green) was crawling in the brain capillary and adopting a rod-shaped morphology when passing through the capillary in CX3CR1 gfp/+ mice 18 h after infection with 20x10 6 C. neoformans. (AVI)