Roles for Microglia in Cryptococcal Brain Dissemination in the Zebrafish Larva

Abstract

Cryptococcal infection begins in the lungs, but yeast cells subsequently access the bloodstream, from which they can reach the central nervous system (CNS). The resulting meningoencephalitis is the most common presentation and is very difficult to treat. How this fungus interacts with the blood-brain barrier (BBB) and establishes growth in the brain parenchyma remains a central question in fungal pathogenesis. We and others have developed the zebrafish larva as a model host for cryptococcosis and demonstrated that hematogenous CNS infection is replicated in this model. Here, we have used this model to examine the details of BBB crossing and the events immediately before and after. We have observed multiple mechanisms of BBB crossing and found that microglia, the resident phagocytes of the brain, likely have multiple roles. First, microglia either actively transfer yeast cells across the BBB or take up a significant proportion of them immediately after crossing. Second, microglia are capable of clearing individual cryptococcal cells at a developmental stage before adaptive immune cells have emerged. Third, microglia serve to maintain endothelial integrity, preventing other, phagocyte-independent forms of crossing. These proposed microglial functions during infection in the zebrafish larva generate new hypotheses concerning the establishment and control of cryptococcal meningoencephalitis. IMPORTANCE Cryptococcal meningitis is a fungal infection of the brain and a major cause of death in people with uncontrolled HIV. Infection begins in the lungs but can enter the bloodstream and disseminate to the brain. A structure called the blood-brain barrier must be crossed for the fungus to enter and cause meningitis. Learning how Cryptococcus crosses the blood-brain barrier will be crucial to understanding and possibly preventing brain infection. Using the zebrafish larva as a model host, we show that microglia, the resident phagocytes of the brain, potentially play multiple previously unappreciated roles in cryptococcal infection of the brain. These roles include reinforcing the integrity of the blood-brain barrier, clearing cryptococcal cells after they have crossed, and possibly participating directly in crossing via a previously unknown mechanism.

Keywords: Cryptococcus; blood-brain barrier; endothelial cells; host-pathogen interactions; macrophages; microglia; pathogenesis.

FIG 1

Quantification of cryptococcal yeast cells in the brain from 1 to 4 dpi. (A) Schematic workflow. Instances of cryptococcal yeast in the brain were quantified in observation and enumerated in the brains of a total of 77 live larvae compiled in 3 replicates. (B) Instances were categorized based on morphology. Note that cells with macrophage morphology are seen both inside and outside the vasculature. (C) Morphological quantification of all brain instances in live larvae. Colors correspond with the bars in B; gray, extracellular; green, macrophages; yellow, microglia; white, uncertain morphology. (D) Quantification of all brain instances in a separate set of fixed larvae using ISH for positive identification of microglia. (E) Quantification of the subset of instances that have crossed the BBB in the live larva set. (F) Quantification of the subset of instances that had crossed in the fixed larva set.

FIG 2

(A to F) Analysis of brain instances in mpr1Δ infection (A and B) and in PU.1 morphant larvae (C to F). (A) Crossing efficiency (number of crossings per instance in the whole experiment) in the original control data set of 77 larvae, matched control (JMD163) infection (closed circle), and mpr1Δ infection (closed square). The latter 2 sets were obtained simultaneously and consisted of 3 replicates of 10 larvae each. Pairwise analyses throughout the figure represent Welch’s t test of log-transformed ratios. (B) Intracellular mpr1Δ yeast cells (brackets) in the parenchyma after crossing the BBB; Ph, uninfected phagocytes. (C) Percent reduction in phagocytes in PU.1 morphants compared to larvae that received control morpholino. All mpeg⁺ cells decreased in the brain, but apoE⁺ cells decreased more. (D and E) Instances per fish (D) and crossing efficiency (E) comparison between the original control data set of 77 larvae, control morpholino (2 replicates of 10 larvae each), and PU.1 morpholino (3 replicates of 16, 15, and 13 larvae). (F) Status of crossing instances in PU.1 morphant infections; n = 3 replicates as in C and D; gray, extracellular; white, intracellular. (G) Example of endothelial breakdown in the brain during infection of PU.1 morphant. Infected Phagocyte inside a deteriorating vessel. White arrowheads indicate fragments of endothelial cytoplasm.

FIG 3

Quantification of mpr1Δ infection of control versus PU.1 morphants. (A) Total brain instances extracellular (gray) and intracellular (white); mpr1Δ + control morphant, n = 3 replicates of 10 larvae each; mpr1Δ + PU.1 morphant, n = 3 replicates of 9, 10, and 10 larvae. (B) Status of crossing instances from same replicates as A. (C) Crossing efficiency in same replicates as A and B. Pairwise analyses throughout the figure represent Welch’s t test of log-transformed ratios.

FIG 4

Macrophage and microglia in BBB crossing and subsequent fate of Cryptococcus. (A) Intravascular macrophage (mΦ) containing cryptococcal yeast cells (cc) extends a pseudopod outside the vasculature and pulls its cargo along with it. Phagocytes (mpeg⁺) and yeast cells express EGFP. Note that nuclear-localized fluorescence signal emphasizes yeast cells versus phagocytes of the same color. Endothelial cells express cytoplasmic mCherry (magenta). (B) A presumed microglial cell (μglia) tightly associated with the abluminal side of a vessel containing Cryptococcus. (C) A cell with microglial morphology in the parenchyma (yellow arrows indicate ramifications) containing yeast cells along with endothelial cytoplasm (white arrowheads). (D) Fate of crossing instances compiled from the original control data set plus additional replicates performed specifically for this analysis. P values indicate results of a Fisher’s exact test applied to contingency data.