Neurons are the Primary Target Cell for the Brain-Tropic Intracellular Parasite Toxoplasma gondii

Abstract

Toxoplasma gondii, a common brain-tropic parasite, is capable of infecting most nucleated cells, including astrocytes and neurons, in vitro. Yet, in vivo, Toxoplasma is primarily found in neurons. In vitro data showing that interferon-γ-stimulated astrocytes, but not neurons, clear intracellular parasites suggest that neurons alone are persistently infected in vivo because they lack the ability to clear intracellular parasites. Here we test this theory by using a novel Toxoplasma-mouse model capable of marking and tracking host cells that directly interact with parasites, even if the interaction is transient. Remarkably, we find that Toxoplasma shows a strong predilection for interacting with neurons throughout CNS infection. This predilection remains in the setting of IFN-γ depletion; infection with parasites resistant to the major mechanism by which murine astrocytes clear parasites; or when directly injecting parasites into the brain. These findings, in combination with prior work, strongly suggest that neurons are not incidentally infected, but rather they are Toxoplasma's primary in vivo target.

Fig 1. Toxoplasma parasites predominantly interact with neurons throughout CNS infection.

Cre-reporter mice were infected with II-Cre or III-Cre Toxoplasma parasites as labeled. Brains were harvested, sectioned, and stained for neurons (anti-neuronal cocktail) and astrocytes (anti-GFAP) at specified time points. Stained sections were analyzed by confocal microscopy to identify if GFP co-localized with stains for neurons, astrocytes, or neither (unidentified). (a) Representative stitched-grid image of a brain section from a III-Cre infected mouse at 3 weeks post infection (wpi). White boxed area in left image is enlarged and separated into the different channels, as labeled (right images). White arrowheads denote GFP⁺ cells that co-localized with anti-neuron staining, red arrowheads denote GFP⁺ cells that did not co-localize with either anti-neuron or anti-astrocyte staining. Left image scale bar, 1 mm. Enlarged image scale bar, 50 μm. (b) Quantification of co-localization for II-Cre infected brain sections at different time points post infection. Bars, mean ±SEM. (c) As in (b) but restricting the analysis only to GFP⁺ cells identified as neurons or astrocytes. (d), (e) As in (b), (c) but for III-Cre infected mice. Bars, mean ±SEM. N = 3–4 mice/time point/Toxoplasma strain. The total number of GFP⁺ cells examined at each time point ranged from 232–372/II-Cre and 368–506/III-Cre. No statistical differences were found in the mean percentage of GFP⁺ neurons across time points in either II-Cre or III-Cre infection (one-way ANOVA, p = 0.13 and p = 0.45 respectively). At 6 wpi, one of the III-Cre infected mice had a substantially lower percentage of GFP⁺ neurons compared to the other mice (67 vs. 91,100,100). Exclusion of this mouse from data analysis changes the mean percentage of GFP⁺ neurons at 6 wpi from 90 ±8 (full data set) to 97 ±3 (1 mouse excluded), which results in a suggestion that in III-Cre infected mice the percentage of GFP⁺ neurons is lower at 1.5 wpi compared to 6 and 12 wpi (one-way ANOVA, p < 0.01). No GFP⁺ cells were found in II-Cre or III-Cre infected brain sections from 0.5 wpi (N = 2 mice/Toxoplasma strain, 9 sections/mouse).

Fig 2. In the setting of IFN-γ depletion, Toxoplasma reactivation leads to an increase in parasite burden, GFP⁺ cells, and GFP⁺ astrocytes in the CNS.

Starting at 4 wpi, isotype control or anti-interferon-γ antibodies were administered every 5 days to Cre reporter mice infected with II-Cre parasites. Mice were sacrificed at 6 wpi after receiving 3 doses of antibody treatment. Brains were sectioned and stained as in Fig 1. Confocal microscopy was used to analyze stained brain sections from control and IFN-γ-depleted mice. (a) Representative stitched-grid image of half of a coronal brain section from a control (left) or IFN-γ-depleted (right) mouse. Scale bar, 1 mm. (b) Quantification of cyst number (6 sections/mouse) and GFP⁺ cell number (1 section/mouse) found in control or IFN-γ-depleted mice. N = 4–5 mice/group. *p< 0.05 by independent sample, two-tailed t-test. (c) Quantification of the lineage of GFP⁺ cells by co-localization with antibody stains for neurons, astrocytes, or neither (unidentified). (d) As in (c) but restricting the analysis only to GFP⁺ cells identified as neurons or astrocytes. N = 4–5 mice/group. N = 82–209 GFP⁺ cells/mouse analyzed for cell lineage studies. ***p< 0.001 by independent sample, two-tailed t-test.

Fig 3. Neuron-parasite interactions predominate during infection with IRG-resistant Toxoplasma parasites.

Cre reporter mice were infected with either III-Cre or III-Cre parasites that express the type I ROP18 protein (III-Cre-ROP18). At 2 weeks post infection (wpi), brains were harvested, sectioned, stained for astrocytes and neurons, and analyzed by confocal microscopy as in Fig 1. (a), (b) Representative merged stitched-grid image of a brain section from a III-Cre (a) or III-Cre-ROP18 (b)-infected mouse. Blue = astrocyte stain (GFAP), Cyan = neuronal stain (neuronal cocktail), Green = GFP expression. Scale bar, 1 mm. (c) Quantification of the lineage of GFP⁺ cells by co-localization with antibody stains for neurons, astrocytes, or neither. N = 100–114 GFP⁺ cells/mouse, N = 5 mice/Toxoplasma strain. (d) As in (c) but restricting the analysis only to GFP⁺ cells identified as neurons or astrocytes. There is no significant difference between the mean percentage of GFP⁺ neurons or astrocytes in III-Cre vs. III-Cre-ROP18 brain sections (p-value = 0.25, independent sample, two-tailed t-test.)

Fig 4. Neuron-parasite interactions dominate in direct CNS inoculation with Toxoplasma, consistent with the baseline ratio of neurons:astrocytes.

Phosphate Buffer Saline (PBS) or Toxoplasma-Cre parasites (II-Cre) were stereotactically injected into the cerebral cortex of naïve Cre reporter mice. At 3, 6, and 9 days post-injection (dpi) brains were harvested, sectioned, stained for astrocytes and neurons, and analyzed by confocal microscopy for co-localization of GFP with astrocyte or neuronal stains or for baseline numbers of astrocytes and neurons. Brain sections from uninfected mice were also stained for neurons and astrocytes and analyzed for baseline numbers of astrocytes and neurons. (a) Schematic of site of injection. (b) Representative merged stitched-grid images of the injection site at 3 dpi from PBS (left) or Toxoplasma-Cre (right) injected mice. Blue = astrocyte stain (anti-GFAP), Cyan = neuronal stain (anti-neuronal cocktail), Green = GFP expression. Scale bar, 200 μm. (c) Quantification of GFP⁺ cells identified as either astrocytes or neurons by co-localization with staining. N = 3 mice/time point, 100 identified GFP⁺ cells/mouse. No statistical differences were found between the mean percentage of GFP⁺ neurons across time points (one-way ANOVA, p = 0.97). No GFP⁺ cells were seen in PBS injected mice. N = 1 mouse/time point. (d) Schematic of fields of view (FOV) taken to assess neuron and astrocyte numbers across IC infection time points. (e) Quantification of number of astrocytes/FOV. (f) Quantification of number of neurons/FOV. As astrocytes in uninfected mice express little GFAP (S4a Fig), these sections were stained with anti-GFAP, anti-S100β, and anti-ALDH1L1 antibodies. The neuronal cocktail was unchanged. UI, uninfected. N = 3–4 mice/time point, 3 randomly selected FOV/hemisphere/ mouse (6 FOV total/mouse) were counted. * p<0.05, *** p< 0.001, ordinary one-way ANOVA.

Fig 5. Toxoplasma cysts are primarily located in neuronal processes.

Cre reporter mice were infected with II-Cre or III-Cre parasites. At 3 weeks post infection, brains were harvested, sectioned into ~200 μm thick sections, processed to render the tissue optically clear, and then imaged at 40x on a confocal microscope. Resulting images were then analyzed with Imaris software to identify the cellular location of the cyst (neuronal process or cell body). For cysts within a process, imaging analysis software was used to determine the distance between the cyst and the cell body, along the infected process. (a) Representative maximal projection image of a cyst within a cell body. S1 Video shows a 3-D movie of this cell. (b) Representative maximal projection image of a cyst within a neuronal process. (c) Representative maximal projection image of (b) analyzed by Imaris to identify the whole neuron and to determine the length from the cyst edge to the edge of the cell body (yellow highlighted line). S2 Video shows a 3-D movie of this cell. Scale bars, 20 μm. (d) Quantification of the percentage of cysts found in the cell body or a neuronal process in mice infected with II-Cre or III-Cre parasites (as labelled). N = 26–28 cysts (III-Cre and II-Cre respectively). (e) Quantification of the distance between cysts and cell bodies for II-Cre and III-Cre cysts within neuronal processes. Cysts in the cell body (N = 7, II-Cre, N = 3, III-Cre) were excluded from this analysis. Each dot represents a single cyst-cell body measurement. Bars, mean ±SEM.