A model for Trojan horse transit into the brain

In the absence of trauma, pathogens that cause lethal brain infections (e.g., those in Table 1) reach it from remote sites, generally traveling in the bloodstream. For microbes that use Trojan horse transit, the first step is infection of a phagocyte in the periphery (Fig 1A). Once internalized, the pathogen may actively manipulate the phagocyte to promote migration towards the brain [7]. Alternatively, it may suppress phagocyte activation (and consequent sequestration in the tissue of origin), allowing the infected cell to circulate normally throughout the body (Fig 1B). Once an infected phagocyte reaches the brain, it adheres to the luminal side of brain capillaries (with or without activation of BMECs) and crosses the BBB, either paracellularly (between BMECs) or transcellularly (through BMECs) (Fig 1C). After brain entry, the pathogen may exit its Trojan horse to infect other neural structures (Fig 1D). This model has been elucidated in the most detail for HIV and other viruses [8, 9], but studies reviewed below suggest that similar strategies are used by other microbes that are the focus of this review: bacteria, fungi, and parasites. Aspects of this model may also apply to non-CNS pathogens, such as mucosal pathogens that use phagocytes to disseminate (see “Perspectives and conclusions”).

Download:

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

Fig 1. The role of phagocytes as Trojan horses for CNS pathogens.

Most neuroinvasive pathogens first infect organs outside the CNS, such as the lungs or the intestines. Cryptococcus neoformans infection of the lungs is shown here as an example. Once infection is established, phagocytes (pink cells) are recruited to these sites (A), where they engulf the pathogen. Some infected phagocytes leave the site of infection and enter the bloodstream, facilitated by the highly permeable vasculature (green) of peripheral organs (B). Through a process that is poorly understood, many of these home to the CNS. Once there, infected phagocytes may act as Trojan horses, traversing the BBB (blue cells) with the pathogen as a passenger (C). Although both paracellular (top) and transcellular (bottom) transmigration can occur, the latter is most likely due to the presence of tight junctions in the BBB (C). Once inside the brain, pathogens can potentially exit their Trojan horses and infect other neural structures (D). Parts of this model (A and B) also apply to phagocyte-assisted dissemination and infection outside of the CNS (see text). BBB, blood–brain barrier; CNS, central nervous system; ECs, endothelial cells. Arrows indicate movement; broken arrow indicates fungal egress.

https://doi.org/10.1371/journal.ppat.1006680.g001

Download:

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

Table 1. Deaths due to CNS infection by select obligate and facultative intracellular microbes.

https://doi.org/10.1371/journal.ppat.1006680.t001

Bacterial infections

The most common causes of bacterial meningitis are the facultative intracellular pathogens Streptococcus pneumoniae and Neisseria meningitidis. Because they survive in the blood and can independently interact with BMECs to enter the brain, these microbes do not require Trojan horse transit, although this mechanism may contribute to S. pneumoniae infection [10]. In contrast, Trojan horses play a central role in infections by another leading cause of bacterial meningitis, Listeria monocytogenes.

L. monocytogenes is a pathogen of humans and domesticated animals that invades the brain parenchyma, unlike most neuroinvasive bacteria, which are limited to the meninges (Table 1). This distinct pathology may relate to its use of Trojan horse invasion, which was first suggested by histological studies showing parasitized phagocytes in the brain tissue of infected mice [11]. Further suggestive of a Trojan horse mechanism were reports that phagocytosis of L. monocytogenes causes the release of immune signals and activation of BMECs [12], both of which would promote the recruitment of additional leukocytes to the site of infection. More direct support for this mechanism came from the observation that infected mice treated with gentamicin to kill extracellular bacteria still developed CNS infection [13]. This occurred regardless of the initial route of infection, consistent with a general model whereby phagocytes are recruited to the site of infection, engulf the pathogen, and then disseminate (Fig 1). Bolstering this observation, injection of L. monocytogenes-infected bone marrow myeloid cells caused faster and greater brain colonization than injection of free bacteria [14]. In this study, which used chimeric mice that expressed a fluorescent protein in their bone marrow cells, an increase in fluorescent signal was observed in the brain as the infection progressed, also consistent with a Trojan horse model. The bone marrow and spleen are among the first organs infected by L. monocytogenes. Interestingly, phagocytes infected in these tissues cannot kill the bacteria but do up-regulate chemokine receptors, making them ideal Trojan horses [15].

Fungal infections

Fungal infections are responsible for up to 1.6 million deaths every year [3], and the ones affecting the CNS have the highest morbidity and mortality [16]. Although several fungal pathogens cause meningitis (Table 1), the only one to frequently do so is Cryptococcus neoformans. Most people have been exposed to this environmental yeast [17]. While healthy individuals are generally asymptomatic, in immunocompromised individuals, the initial pulmonary infection can subsequently disseminate to the CNS. As a result, C. neoformans is the most common causative agent of meningitis in sub-Saharan Africa and a leading cause of death in HIV+ individuals, killing close to 200,000 people each year [18].

As with L. monocytogenes, early evidence for Trojan horse transit of C. neoformans came from histological examination of brains from infected mice [19]. This work was complemented by studies supporting the role of phagocytes in cryptococcal dissemination from the lungs to the brain. For example, depletion of alveolar macrophages reduced dissemination from the lungs [20], systemic monocyte depletion after lung infection reduced fungal burden in other organs [21], and intravenous administration of C. neoformans-associated macrophages caused higher brain burden than infection with free cryptococci. More recently, direct evidence for Trojan horse transit has come from two studies using in vitro models of brain endothelia. Both groups cultured human cerebral microvascular endothelial cell (hCMEC) monolayers on permeable membranes separating the upper (“blood”) and lower (“brain") compartments of tissue culture wells. In one study, a monocytic cell line was first incubated with C. neoformans, which was engulfed by or adhered to the phagocytes, and the samples were then stained to mark any externally adherent fungi. This mixture was added to the upper chamber, and one day later, monocytes containing unstained fungi were found in the lower chamber, suggesting that Trojan horse crossing had occurred [22]. (Interestingly, the same experiments performed with Cryptococcus gattii, a species that primarily causes lung infections, showed less barrier crossing.) In the other study, our group used a flow cytometry strategy to isolate primary human monocytes or macrophages that contained only a single internalized cryptococcal cell. We used this population to directly compare Trojan horse and free fungal transit across a similar BBB model and found that both mechanisms contribute to overall transmigration [23]. We further showed that immune signals that are normally generated during cryptococcal infection preferentially stimulate Trojan horse transit and that this mode of entry provides an alternative for fungal mutants that cannot otherwise traverse the BBB. Finally, we used live microscopy to directly visualize C. neoformans-infected phagocytes as they crossed model BBB by forming transendothelial pores in the hCMEC. Our microscopic observations also suggested that phagocytes may serve as “taxis” in addition to Trojan horses, contributing to brain infection by picking up the cryptococci (which survive poorly in blood) at distal sites and delivering them to the BBB, where the free fungi can cross independently.

Parasitic infections

Parasitic infections cause high burdens of disease in low- and middle-income countries, with almost 800,000 deaths in 2015 (Table 1). Several parasites cause devastating CNS pathology, either while remaining in the vasculature—like the parasite that causes malaria—or by crossing the BBB. Here we focus our discussion on a parasite that is estimated to infect one-third of the world, Toxoplasma gondii [24].

T. gondii is acquired orally and colonizes the gastrointestinal tract. In healthy humans, a robust immune response halts the rapid parasite proliferation that would cause severe acute disease in an immunocompromised host. However, even immunocompetent individuals do not completely clear the infection and remain chronically infected with quiescent parasite cysts, mainly in tissues of the CNS and skeletal muscle. Support for Trojan horse transport of T. gondii derives from studies similar to those mentioned above for other pathogens, mainly adoptive transfer studies showing that parasitized monocytes or dendritic cells cause brain infection faster than free parasites [25]. Consistent with these observations, the injection of intracellular parasites together with antibodies against CD11b, which blocks phagocyte migration, reduced brain burden 2-fold. Furthermore, enhanced transendothelial migration of infected leukocytes has been observed in some, although not all, in vitro studies [26, 27]; another study using a robust BBB model consisting of brain endothelia and astrocytes reported the preferential transmigration of infected monocytes [28]. Lastly, T. gondii Trojan horse transit has been visualized in vitro, although these studies used an activated non-brain endothelial cell line (human umbilical vein endothelial cells [HUVECs])[29].

As with the other pathogens discussed here, free T. gondii likely also cross the BBB, although the relative frequency of the two processes is not known. Notably, infection causes endothelial cells to become activated, with up-regulation of adhesion molecules and down-regulation of junctional complexes [28]; both of these processes could stimulate phagocyte transmigration and thus promote Trojan horse transit. Intravital microscopy has also shown that BMECs serve as a replicative niche for T. gondii and that intracellular replication is required for egress (through host cell lysis) into the CNS [30]. Interestingly, the same experiments did not show Trojan horse movement, although they did reveal infected phagocytes trapped on the vascular side of brain vessels; these may act as taxis (as with C. neoformans), serving as a source of free parasites to infect BMECs or cross the BBB. This idea has recently been supported by the observation that adhesion of infected leukocytes to endothelial cells in vivo triggers parasite egress [31].