A noncanonical autophagy is involved in the transfer of Plasmodium-microvesicles to astrocytes

ABSTRACT

Cerebral malaria is a neuroinflammatory disease induced by P. falciparum infection. In animal models, the neuro-pathophysiology of cerebral malaria results from the sequestration of infected red blood cells (iRBCs) in microvessels that promotes the activation of glial cells in the brain. This activation provokes an exacerbated inflammatory response characterized by the secretion of proinflammatory cytokines and chemokines, leading to brain infiltration by pathogenic CD8⁺ T lymphocytes. Astrocytes are a major subtype of brain glial cells that play an important role in maintaining the homeostasis of the central nervous system, the integrity of the brain–blood barrier and in mounting local innate immune responses. We have previously shown that parasitic microvesicles (PbA-MVs) are transferred from iRBCs to astrocytes. The present study shows that an unconventional LC3-mediated autophagy pathway independent of ULK1 is involved in the transfer and degradation of PbA-MVs inside the astrocytes. We further demonstrate that inhibition of the autophagy process by treatment with 3-methyladenine blocks the transfer of PbA-MVs, which remain localized in the astrocytic cell membrane and are not internalized. Moreover, bafilomycin A₁, another drug against autophagy promotes the accumulation of PbA-MVs inside the astrocytes by inhibiting the fusion with lysosomes, and prevents ECM in mice infected with PbA. Finally, we establish that RUBCN/rubicon or ATG5 silencing impede astrocyte production in CCL2 and CXCL10 chemokines induced by PbA stimulation. Altogether, our data suggest that a non-canonical autophagy-lysosomal pathway may play a key role in cerebral malaria through regulation of brain neuro-inflammation by astrocytes.

Introduction

Cerebral malaria (CM) is a life-threatening disease induced by Plasmodium falciparum infection in humans. It occurs principally in children and immunocompromised individuals, and results in approximately 405,000 deaths per year globally [1]. The pathophysiology of CM is complex and multifactorial, and many aspects of the disease remain poorly understood. Apart from the occurrence of numerous metabolic, pathological, and physiological abnormalities caused by a dysregulated proinflammatory response, autoantibodies against brain components are also associated with CM development [2,3]. Experimental CM (ECM) induced by infecting C57/BL6 mice with the P. berghei ANKA (PbA) strain is a well-established model that mimics many characteristics of human CM. In mice developing ECM, the concomitant presence of brain-sequestrated parasitized erythrocytes, αβ CD8⁺ T cells, and CD4⁺ IL2RA/CD25⁺ FOXP3⁺ regulatory T cells has been clearly demonstrated [4–6]. The neuropathogenesis of ECM involves a detrimental immune response characterized by exacerbated neuro-inflammation. Distinct but complementary mechanisms have been shown to activate astrocytes and microglia, the two important immune cell types of brain parenchyma [7,8]. This results in elevated production of proinflammatory cytokines such as IFNG/IFN-γ (interferon gamma) and TNF/TNF-α (tumor necrosis factor) and chemokines such as CCL2 (chemokine (C-C motif) ligand 2), CXCL10 (chemokine (C-X-C motif) ligand 10), and CCL3/MIP-1α (chemokine (C-C motif) ligand 3). In turn, this pro-inflammatory response promotes the infiltration of lymphoid cells into the brain, edema, axonal damage, and myelin loss [4,9].

Astrocytes are one of the most prominent glial cells of brain parenchyma [10]. They constitute a heterogeneous cell population that plays an important role in brain homeostasis, neuronal functions, and the maintenance of the integrity and selectivity of the brain–blood barrier (BBB) [11,12]. Astrocytes are a part of the brain’s innate immune system that actively participates in various neuropathological disorders during infections and/or inflammatory processes, including CM [11,13]. A marked activation of glial cells has been reported in CM patients [14]. In the PbA model of ECM, we as well as others have shown that astrocytes, together with microglia, exacerbate neuro-inflammation. They are responsible for the high levels of proinflammatory chemokines and cytokines, such as TNF/TNF-α, CXCL10, CCL2, and IFNG/IFN-γ, observed in ECM [7,8,15].

We have recently described the intracellular transfer of P. berghei ANKA-microvesicles (PbA-MVs) from infected red blood cells (iRBCs) to astrocytes early in the infection after contact is established between these two cell populations. This promotes a rapid and strong proinflammatory response that results in the release of CCL3/MIP-1α and CXCL10 [15], thereby leading to BBB disruption and the recruitment of pathogenic CD8⁺ T cells to the brain. Interestingly, the transcriptome of astrocytes and microglia upon contact with iRBCs shows increased activation of genes involved in the intracellular degradation pathway, suggesting a potential role of autophagy-lysosomal pathway in the degradation of transferred PbA-MVs [15].

Autophagy is a self-degradation process involving a spatially and temporally orchestrated machinery, which is activated in response to stress, and which is present in all mammalian cells and tissues, including those of the central nervous system [16,17]. It is a highly conserved process that directs unnecessary or damaged intracellular material to the lysosomes to digest and recycle all types of macromolecules [18]. The initiation of the autophagy process involves a network of ATG (autophagy related) genes and proteins that promote the formation of MAP1LC3/LC3 (microtubule-associated protein 1 light chain 3)-positive autophagosome precursors termed phagophores. The process starts with the engulfment of intracellular cargo by the phagophore. The phagophore then expands through the acquisition of lipids into a double-membrane structure termed the autophagosome [17,19]. The autophagosome subsequently undergoes a process of maturation, and it eventually fuses with the lysosome to form the autolysosome, which has the ability to degrade substrates to produce amino acids and other metabolites [19–21]. During infection by intracellular pathogens such as Plasmodium, two types of autophagy pathways can be engaged: (a) nonselective autophagy, which is as an adaptive response for cellular remodeling upon unfavorable stress conditions that can also support parasite nourishment and (b) selective autophagy, which specifically targets intracellular parasites and acts through ubiquitination for their elimination [22–25].

Selective autophagy is initiated by ULK1 (unc-51 like kinase 1) following the inhibition of MTOR (mechanistic target of rapamycin kinase), which then drives the formation of a phagophore by activating and recruiting the complex BECN1 (beclin 1, autophagy related) and ATG proteins (ATG12–ATG5-ATG16L1) [17,26,27]. LC3-I is then converted to LC3-II (the form conjugated to phosphatidylethanolamine) and attached to the phagophore membrane to package the cargo with the help of receptors such as CALCOCO2/NDP52 (calcium binding and coiled-coil domain 2) or SQSTM1/p62 (sequestosome 1) [28–30]. The final step in the degradation of the cargo takes place near the nucleus, with the fusion of the autophagosome with the lysosome [17]. However, MAP1LC3/LC3 and some ATG proteins are also involved in the noncanonical autophagy process, the LC3-associated phagocytosis (LAP), which is a distinct pathway that functions independently of ULK1 and which induces a LAPosome bound by a single membrane [31].

In this study, we investigate whether autophagy is associated with PbA-MVs transfer to astrocytes. We present evidence that autophagy-lysosomal pathways are involved in the uptake of PbA-MVs by primary cultures of astrocytes after interaction with PbA-iRBCs expressing green fluorescent protein (GFP). Moreover, we investigate the links between the autophagy process and proinflammatory responses induced in astrocytes that lead to ECM. Our results demonstrate a key role of an unconventional ULK1-independent LAP pathway in the transfer of PbA-MVs. Upon interaction of PbA-iRBCs with astrocytes, microvesicles are formed and transferred to these important brain immune cells, and this process likely induces neuro-inflammation associated with parasite infection and ECM.

Results

Microvesicles are transferred in astrocytes upon contact with PbA-iRBCs

In order to determine the kinetics of microvesicle biogenesis in astrocytes in contact with malarial parasite-infected RBCs, we incubated primary astrocyte cultures derived from newborn C57BL/6 mice with GFP-labeled PbA-iRBCs at ratio 1:10. Scanning confocal microscopy revealed GFP and DAPI stained microvesicles inside the astrocytes upon 6 h of contact (Figure 1A). For confirmation that these GFP-PbA-MVs contained parasite material, they were stained for the parasite protein UIS4 (Figure 1A, B). We found that GFP-PbA-MVs were localized inside the astrocyte cytoplasm and surrounding parasitophorous vacuole membrane (PVM) expressing UIS4 (Figure 1A, B). The presence of parasite material inside astrocytes was also confirmed by the quantification of Pb18s (Plasmodium berghei 18S ribosomal gene) – a ribosomal subunit gene of PbA. A significant increase in the amount of Pb18s was observed 6 h post-co-culture (Figure 1C; ***P < 0.001) that decreased after 24 h (***P < 0.001) and reverted to control levels after 48 h (Figure 1C; ***P < 0.001). The decrease in the relative amount of Pb18s after 24 h strongly suggested a parasite degradation. These data also indicated that GFP-PbA-MVs transferred inside astrocytes upon PbA-iRBCs contact were derived from the parasite and disappeared with time as shown at 24 and 48 h post contact.

Figure 1.

PbA-MVs expressing UIS4 and parasite material are observed inside astrocytes that are in contact with PbA-iRBCs for 6 h. Confocal microscopy and RT-qPCR analyses of primary astrocyte cultures incubated with GFP-PbA-iRBCs at a ratio of 1:10 for either 1) 3 h, or 2) 6 h, or 3) PbA-iRBCs were removed after 6 h, and the astrocytes were incubated further for 24 or 48 h. (A) Confocal microscopy of 6 h-stimulated cells stained with DAPI for the DNA (blue) and with goat anti-UIS4 antibodies for the PVM (magenta). The orthogonal views X-Y and Y-Z generating 3D data image of PbA-MVs enclosed by the parasitophorous vacuole are shown. (B) The same orthogonal views are illustrated in split channels (representative of three experiments). (C) Relative amount of Pb18s gene quantified by RT-qPCR (n = 5 independent cell cultures per group). Data represents median fold-change ± SEM of Pb18s gene expression at different time points after GFP-PbA-iRBCs stimulation as compared to unstimulated controls. Statistical analysis was done using one-way ANOVA with Tukey’s multiple comparison test, and data were considered significant at *P < 0.05; **P < 0.01; ****P < 0.0001.

The LC3-mediated autophagy pathway is involved in the transport of parasite microvesicles in the astrocytes

A transcriptomic analysis of astrocyte cultures after 6 h of GFP-PbA-iRBCs contact revealed an increased expression of intracellular degradation pathway genes [15]. We therefore investigated whether autophagy, an intracellular recycling pathway of eukaryotic cells, is involved in the transfer of PbA-MVs inside astrocytes. We measured MAP1LC3/LC3 expression on PbA-MVs in astrocytes at different time points following their interaction with GFP-PbA-iRBCs by confocal microscopy using anti- MAP1LC3/LC3 antibodies and DAPI staining. (Figure 2A, Figure 2B) shows DAPI stained GFP-PbA-MVs in astrocytes abundantly surrounded by MAP1LC3/LC3 proteins at 6 h of stimulation, as compared to unstimulated control. The orthogonal views (X-Z and Y-Z) confirmed the intracellular position of MAP1LC3/LC3⁺-PbA-MVs (Z = 5) (Figure 2B). Confocal imaging for GFP and MAP1LC3/LC3 co-expression at 24 and 48 h after PbA-iRBCs removal showed increased accumulation of PbA-MVs localized around the nucleus (Z = 3), and a diffusion of the green fluorescence into the cytoplasm (Figure 2C). An increase in the number of PbA-MVs-containing astrocytes was observed with increasing time. Twenty-six percent (26.06 ± 2.15%) of the astrocytes were GFP⁺ after 3 h of contact, peaking at 41.92 ± 1.46% after 24 h and remained at this level after 48 h (Figure 2B, C). To clarify the association between GFP-PbA-MV and MAP1LC3/LC3 expression in astrocytes, we performed an interactive microscopy image analysis using the Imaris software. Automatic detection of intracellular GFP-PbA-MV and red- MAP1LC3/LC3 fluorescent spots of 2- to 5-µm size was done on total astrocytes for each confocal microscopy slide after cell segmentation by delimitation of the cell membrane to focus on intracellular events (Figure 2D). The data showed a significant increase of GFP⁺-PbA-green (***P < 0.001), concomitantly to MAP1LC3/LC3⁺-red fluorescent spots (**P < 0.01) at 6 h inside astrocyte that decrease at 48 h (**P < 0.01) (Figure 2D). The presence of GFP signal inside astrocytes and its localization into structures that resembled phagophores expressing MAP1LC3/LC3⁺ at 6 h post contact suggested the involvement of an autophagy mechanism (Figure 2). These observations were supported by TEM analysis revealing the presence of PbA-containing phagophores within the astrocytes 24 h after contact (Figure 3A, white arrows). MAP1LC3/LC3⁺ PbA-MVs-containing phagophores were probably localized in the area corresponding to lysosomes near the nucleus, the place known to harbor the final step of the autophagy process [22,29]. To assess phagophores and lysosome fusion, we monitored the lysosomal trafficking in GFP-PbA-iRBC stimulated astrocytes by real-time microscopy. Figure 3B shows the green fluorescence of two GFP-PbA-containing phagophores located next to the nucleus (at 9.40 and 12 h, respectively). These phagophores colocalized with the red fluorescence of LysoTracker Red and followed the same trend of decreasing intensity in time at 10.40 and 12.20 h, respectively (Figure 3B, C). Movie S1 shows a dynamic event corresponding to the lysosomal activity and trafficking in astrocytes that had been in contact with GFP-PbA. The late phase of lysosomal activity of the phagolysosome is characterized by the presence of black aggregates. TEM revealed the presence of such structures near the nuclei of astrocytes (Figure 3D). To further characterize the autophagy process involved in the transfer of PbA-MVs, we quantified the expression of several genes of the pathway, including Becn1, Map1lc3/Lc3, Atg5, Atg16l1, Sqstm1/p62, and Calcoco2/Ndp52, and compared them to unstimulated and cells stimulated with uninfected RBCs using RT-qPCR (Figure 4A and Fig. S1). A significant decrease in the expression of Becn1 was observed between 6 and 24 h post astrocyte-PbA-iRBCs contact, followed by a significant upregulation at 48 h (**P < 0.01) (Figure 4A). By contrast, the expression of Map1lc3/Lc3, Atg5, and Atg16l1 genes, known to be required for autophagosome formation, was increased during the 6 h of contact but was significantly decreased at 24 h (Map1lc3/Lc3 ***P < 0.001, Atg5 ***P < 0.001, Atg16l1 **P < 0.01) and 48 h (Atg5 and Atg16l1 **P < 0.01; Map1lc3/Lc3 ***P < 0.001) (Figure 4A). Interestingly, Calcoco2/Ndp52 was significantly downregulated in PbA-iRBCs stimulated astrocytes whereas Sqstm1/p62 decreased significantly after 6 h, suggesting potential involvement of the SQSTM1/p62 receptor in the initial stages of the autophagy process (Figure 4A). Using western blots to assess expression of the ATG16L1, SQSTM1/p62, and BECN1 proteins, we found increased expression of these proteins in astrocytes that had interacted with PbA-iRBCs as compared to controls (0 h) and, except for in the case of SQSTM1/p62, this increased expression continued for 48 h (Figure 4B, Table S1). Expression of SQSTM1/p62 declined at 24 h but rebounded at 48 h (Figure 4B, Table S1). Interestingly, LC3-I expression diminished significantly with time (3–48 h), whereas LC3-II expression was highly variable, although there seemed to be a trend toward increased expression at 6 h (Figure 4B, C). These data strongly suggest that LC3-I to LC3-II conversion was induced in astrocytes upon PbA-iRBCs contact and triggered an autophagy process that may led to PbA-MVs transfer and parasite destruction by the astrocytes.

Figure 2.

PbA-MVs inside astrocyte are MAP1LC3/LC3⁺. Primary astrocyte cultures incubated with GFP-PbA-iRBCs at a ratio 1:10 during 3 h and control cultures were examined after washing at different time point from 3, 6, 24, and to 48 h post incubation. (A) control cultures, (B) the GFP-PbA-MVs (green) containing DNA stained with DAPI (blue) and surrounded by the autophagosomal membrane stained with anti-MAP1LC3/LC3 antibodies (red) (n = 5). The extreme right panel shows the orthogonal views X-Z and Y-Z generating 3D data picture of PbA-MVs into an MAP1LC3/LC3⁺autophagosome, whereas the MAP1LC3/LC3 panel shows the diameter of the MAP1LC3/LC3⁺ autophagosome. (C) Diffusion of GFP fluorescence at 24 and 48 h postincubation. (D) Intracellular GFP⁺-PbA-green fluorescent spots and MAP1LC3/LC3⁺-red fluorescent spots on total cells. Data is represented as mean ± SEM (n = 3). Statistical significance was determined using one-way ANOVA with Tukey’s multiple comparison test. P values indicate statistical significance as at *P < 0.05; **P < 0.01; ***P < 0.001.

Figure 3.

Kinetics of lysosomal fusion with autophagosomes containing GFP-PbA-MVs in astrocytes. (A) Primary astrocyte cultures were incubated with GFP-PbA-iRBCs at a ratio of 1:10 for 6 h and followed for 24 h. TEM micrographs of two independent experiments showing autophagosome containing one PbA-MV (AP) attached to intra-astrocyte lysosome (L) 24 h after contact with PbA-iRBCs. White arrows show membrane contact of the two organelles. (B) Primary astrocyte cultures were incubated with GFP-PbA-iRBCs (green) for 4 h, and lysosomal activity in astrocytes was followed with live confocal microscopy for 19 h using LysoTracker Red (red). Each arrow indicates a fusion event (spot 1 upper panels; spot 2 lower panels). Images are representative of three independent experiments. (C) Quantification of LysoTracker Red mean fluorescence intensity of spots 1 and 2 over time. (D) A TEM micrograph of PbA-MV degradation (diffused black matter) within an autolysosome (AL) located near the nucleus (N) 24 h after contact of the astrocyte with GFP-PbA-iRBCs (representative of three independent experiments).

Figure 4.

Autophagy genes and proteins are expressed in astrocytes after contact with GFP-PbA-iRBCs. Primary astrocyte cultures were incubated with GFP-PbA-iRBCs at a ratio of 1:10 for 3 or 6 h and followed for 48 h and assessed for gene expression by RT-qPCR and protein expression by western blots. (A) Relative expression of autophagy-related genes Becn1, Map1lc3/Lc3, Atg5, Atg16l1, Sqstm1/p62, and Calcoco2/Ndp52 as compared to that of unstimulated cells. Data are median fold-change ± SEM (n = 5 independent experiments per group). Values were compared with a control group (0 h) and intergroup using a two-way ANOVA with Bonferroni post-hoc test, with significance indicated as *P < 0.05; **P < 0.01; ***P < 0.001. (B, C) Kinetics of expression of autophagy-related proteins ATG16L1, SQSTM1/p62, BECN1, LC3-I, and LC3-II. (B) A representative western blot, and (C) Relative fold-change in LC3-I and LC3-II expression in stimulated cells quantified by western blot when compared to unstimulated controls (0 h). Protein content was quantified using ImageJ software (n = 3 independent experiments). One-way ANOVA with Tukey’s multiple comparison test was used for statistical analysis with significance indicated as ***P < 0.001.

A noncanonical autophagy machinery independent of ULK1 is involved in the transfer of PbA-MVs in astrocytes

Next, we investigated the morphology of the autophagosome by TEM that revealed a single-membrane LAPosome packed with the Plasmodium pigment hemozoin when astrocytes were incubated with PbA-iRBCs (Figure 5A). These observations support the notion that a noncanonical LAP pathway may be induced in astrocyte after the transfer of PbA-MVs. It is interesting to note that the LAPosome contains a single-membrane whereas the canonical autophagy is characterized by a doubled-membrane autophagosome. We further used RNA silencing strategy by targeting those genes involved in the classical pathway, Ulk1 and Rb1cc1 (RB1-inducible coiled-coil 1), and, Atg5 and Rubcn/Rubicon (RUN domain and cysteine-rich domain containing, Beclin 1-interacting protein), which are known to be involved in the noncanonical autophagy process using astrocytes after PbA-iRBCs contact (Figure 5B). In other words, we switched off Ulk1, Rb1cc1, Atg5 and Rubcn/Rubicon genes in astrocytes after contact with PbA-iRBCs and quantified by RT-qPCR expression of Map1lc3/Lc3, Atg5 and Atg16l1 genes (Figure 5C). While no difference is observed in siUlk1 and siRb1cc1 conditions compared to stimulated astrocytes at 24 h, Map1lc3/Lc3, Atg5, and Atg16l1 genes are significantly downexpressed in siRubcn/Rubicon and siAtg5 transfected astrocytes compared to the stimulated cells (Figure 5C). However, no difference is observed when RUBCN/rubicon and ATG5 were silenced (Figure 5C). Using RT-qPCR, we found that the expression of Ulk1 gene encoding a kinase that is specifically involved in canonical autophagy was downregulated 24 h post contact with PbA-iRBCs (Figure 5D). Notably, the decrease in Ulk1 gene expression was similar to that observed in rapamycin-treated astrocyte cultures, an activator of autophagy by inhibition of MTOR (Figure 5D). By contrast, a significant increase in Ulk1 gene expression was observed in cultures treated with 3-MA, an inhibitor of autophagy targeting conversion of LC3-I into LC3-II (Figure 5D). Altogether, these data established that LAP, a noncanonical autophagy pathway independent of ULK1 was involved in PbA-MVs transfer and degradation within astrocytes.

Figure 5.

ULK1 independent autophagy process is involved in the LAPosomal degradation of PbA-MVs in astrocytes. Primary astrocyte cultures were incubated with GFP-PbA-iRBCs at a ratio of 1:10 for 6 h, PbA-iRBCs were removed after 6 h, and astrocytes were incubated further for 24 h. The cultures were transfected or not with 25 nM siUlk1, siRb1cc1, siRubcn/Rubicon or siAtg5 RNA; or treated or not with 10 mM 3–MA or 5 mM rapamycin from the time of contact to 24 h. (A) TEM micrograph showing a LAPosome packaging one PbA-MV surrounded by a single limiting membrane within untreated astrocyte 24 h after PbA-iRBCs contact (PbA-MV at 24 h). (B) Checkpoints of ULK1, RB1CC1, RUBCN/rubicon, and ATG5 RNA silencing by RT-qPCR in astrocytes control or stimulated with RBC or iRBCs, with or without siRNA. (C) Quantification by RT-qPCR of relative Becn1, Map1lc3/Lc3, Atg5, and Atg16l1 genes expression in untreated, siUlk1-, siRb1cc1-, siRubcn/Rubicon-, and siAtg5- transfected astrocytes at 24 h after stimulation by PbA-iRBCs (n = 5 per group). (D) Relative Ulk1 gene expression by untreated, rapamycin-, and 3-MA-treated astrocytes quantified by RT-qPCR (n = 5 per group) at 24 h. Data are median fold-change ± SEM of genes expressed. One-way ANOVA with Tukey’s multiple comparison test was used for statistical analysis, with data considered significant at *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Autophagy inhibitors impaired the transfer and the degradation of PbA-MVs in astrocytes

We further analyzed the effects of autophagy activator rapamycin, autophagosome inhibitor 3-MA, and lysosomal fusion inhibitor BAF A₁ on PbA-MVs transfer to astrocytes. Rapamycin targets MTOR, and MTOR inhibition is known to activate autophagy; 3-MA blocks LC3-I to LC3-II conversion involved in the autophagosome elongation and BAF A₁ impairs the autophagosome and lysosome fusion. Astrocytes were treated either with 5 mM rapamycin, 10 mM 3-MA or 10 nM BAF A₁ at time ‘0ʹ, i.e, at the time of addition of GFP-PbA-iRBCs, and the drugs were maintained in culture for 48 h. We then analyzed the fate of PbA-MVs using confocal imaging and quantified the fluorescence after staining the DNA with DAPI, and GFP and MAP1LC3/LC3 with appropriate fluorescent labeled-antibodies. Rapamycin treatment resulted in GFP⁺ PbA-MVs, but with a diffused fluorescence, containing DAPI-labeled DNA inside MAP1LC3/LC3-positive vesicles in the astrocytes (Figure 6A). These PbA-MVs were located close to the nuclei of astrocytes (Figure 6A). In contrast to rapamycin, astrocytes (DAPI in blue at Z = 1) treated with 3-MA showed PbA-MVs (stained blue with DAPI and showing high GFP green staining) that were localized to the plasma membrane of the astrocytes and not enter the cells (Z = 4) (Figure 6A). Red- MAP1LC3/LC3⁺- and GFP⁺-PbA fluorescent spots were significantly increase at 3 h in astrocyte treated with rapamycin compared to 3-MA (Figure 6B) or untreated cells (Figure 2D). After 3 h, we observed a significant reduction overtime of red-MAP1LC3/LC3⁺ and GFP⁺-PbA fluorescence and particularly at 24 and 48 h (Figure 6B). Moreover, 3-MA treatment resulted in a significant reduction in the amount of Pb18s gene in astrocytes when compared to rapamycin or no treatment (Figure 6C; *P < 0.05), suggesting that 3-MA induced the inhibition of LAPosome formation and thereby caused the accumulation of PbA-MVs at the plasma membrane, after 3 h of treatment.

Figure 6.

Effects of inhibition and activation of autophagy on GFP-PbA-MVs transfer in astrocytes. Primary astrocyte cultures were incubated with GFP-PbA-iRBCs at a ratio of 1:10 for 6 h, PbA-iRBCs were removed after 6 h, and astrocytes were incubated further for 48 h. The cells were treated or not with 5 mM rapamycin or 10 mM 3-MA from 0 to 48 h of contact. (A) The presence of MAP1LC3/LC3⁺-LAPosomes (red) enclosing DNA-containing (blue) GFP-PbA-MVs (green) in rapamycin-treated astrocytes, compared 3-MA treated culture, after 6 h of GFP-PbA-iRBCs contact as analyzed by confocal microscopy. The orthogonal view Y-Z of GFP-PbA-MVs within the intracellular compartment is shown. (B) Intracellular GFP⁺-PbA-spots and red-MAP1LC3/LC3⁺-spots per astrocyte, in rapamycin and 3-MA treated astrocyte cultures. (C) Relative expression of Pb18s gene in 3-MA- and rapamycin-treated stimulated astrocyte cultures quantified by RT-qPCR. Data are representative of four independents experiments. Data are mean ± SEM (n = 3) for spots quantification and median fold-change ± SEM (n = 5) for gene expression. Values were compared with control groups (0 h or untreated) using one-way ANOVA with Tukey’s multiple comparison test for spots quantification comparison and Student’s t-test for gene expression comparison, with significant difference indicated as *P < 0.05; **P < 0.01; ***P < 0.001.

Then, we analyzed the expression of autophagy markers in the drug-treated astrocyte-PbA-iRBCs cultures using RT-qPCR and immunoblots. There was an overall decrease in the expression of Becn1, Atg5, and Sqstm1/p62 genes in 3-MA-treated cells when compared to untreated controls (Fig. S2). There was also a decreased expression of Map1lc3/Lc3 and Calcoco2/Ndp52 at 24 h (**P < 0.01; ***P < 0.001 respectively) and 48 h (***P < 0.001; *P < 0.05 respectively) in the treated cultures (Fig. S2). By contrast, expression of Atg16l1 was upregulated at 3 h (***P < 0.001) and 6 h (**P < 0.01) (Fig. S2). The fact that Atg16l1 levels were increased in 3-MA-treated astrocytes stimulated with PbA-iRBCs reinforced our hypothesis that an autophagy process was involved in the PbA-MVs transfer. Furthermore, Western blots revealed decreased autophagic flux (ratio LC3-II:LC3-I) by astrocytes after 3 h post-incubation. However, this ratio LC3-II:LC3-I seems to increase after 24 h post-incubation with PbA-iRBCs in 3-MA treated cultures (Fig. S3). In contrast to 3-MA treatment, we observed significant inhibition of gene expression of all autophagy markers tested at 48 h post-PbA-iRBCs stimulation in rapamycin treated cultures (Fig. S4). Expression levels of Becn1, Map1lc3/Lc3, Atg16l1, and Sqstm1/p62 genes were not inhibited at 3 h, whereas levels of Atg5 and Calcoco2/Ndp52 were inhibited even at this timepoint (Fig. S4). Although Map1lc3/Lc3 gene expression was downregulated, the autophagic flux (ratio LC3-II:LC3-I) was found to be increased in rapamycin-treated cells at 3 and 6 h after astrocyte-PbA-iRBCs interaction (Fig. S3). Our data suggest an inhibition of PbA-MVs transfer and blockage at the plasma membrane in 3-MA treated astrocyte cultures. In contrast, rapamycin treatment seemed to promote the early transfer of PbA-MVs in the astrocytes and their subsequent degradation by autophagy. Together, these observations indicated that autophagy was likely necessary for the transfer and the degradation of PbA-MVs in astrocytes. Next, we monitored the LAPosome-lysosome fusion during astrocyte-PbA-iRBCs interaction by real-time microscopy using LysoTracker Red (red-fluorescence) in culture treated or not with BAF A₁. Figure 7A shows no accumulation of LysoTracker Red colocalized with GFP PbA-MVs fluorescence (spots 1 and 2), in BAF A₁ treated astrocytes (Figure 7A). Then the expression of LC3-II and LC3-I was quantified by Western blot analysis (Figure 7B). We observed an increase in the autophagic flux (ratio LC3-II:LC3-I) in BAF A₁-treated astrocytes compared to untreated. Altogether, these data reinforced our hypothesis of an involvement of a LAP mechanism in the degradation of PbA-MVs in the astrocytes.

Figure 7.

BAF A₁ treatment induces an accumulation of GFP-PbA-MVs inside astrocytes. (A) Primary astrocyte cultures were incubated with GFP-PbA-iRBCs (green) for 4 h and treated or not with 10 nM of BAF A₁. The lysosomal activity in astrocytes was followed with live confocal microscopy for 19 h using LysoTracker Red (red). Each arrow indicates a GFP-LysoTracker Red colocalization event in spot 1 and spot 2 representative of GFP-LAPosomes accumulated in astrocyte). Data are representative of three independent experiments. (B) The autophagic flux (ratio LC3-II:LC3-I) in BAF A₁-treated astrocytes, compared to the control unstimulated (0 h). Data are representative of three independent experiments.

LAP is involved in the development of cerebral malaria

To demonstrate the physiological relevance of LAP in the disease, we evaluated the impact of in vivo treatment with BAF A₁ on the development of ECM induced in susceptible C57BL/6 mice infected with 10⁶ GFP-PbA-iRBCs. Treatment with 2 µg BAF A₁ resulted in a 100% survival of PbA-infected mice (Figure 8A), without affecting the parasitemia. These drug-treated and infected mice died later on from hyperparasitaemia at 23 dpi (>45%) (Figure 8B) compared to untreated mice that died from a neurological syndrome at 6.5 dpi (Figure 8A) with lower parasitemia rate (<7%) (Figure 8B). BAF A₁ is a potent inhibitor of the LAP mechanism involved in the degradation of PbA-MVs in the astrocyte. Protection against ECM suggests its participation in the neuropathological mechanisms.

Figure 8.

BAF A₁ prevented ECM. (A) Survival and (B) Parasitemia of mice infected with PbA and treated with 0.1 mg/kg BAF A₁ from 3 to 5 dpi. Data are results from two independent experiments (n = 5 mice/group) that were compared using the log-Rank (Mantel-Cox) test. Statistical significance **P < 0.01.

LAP is necessary for the induction of PbA-induced proinflammatory response in astrocytes

We then investigated the temporal profiles of cytokines and chemokines produced by primary astrocyte cultures in response to contact with PbA-iRBCs using ELISA. We found that the levels of CXCL10 and CCL2 in culture supernatants increased after 6 h of PbA-iRBCs contact and continued to increase until 48 h. By contrast, the levels of TNF/TNF-α rose within 3 h of PbA-iRBCs contact and, except for a dip at 24 h, continued to increase until 48 h (Figure 9A). Interestingly, IL1B/IL1β (interleukin 1 beta) levels were essentially unaltered, whereas IL6 (interleukin 6) secretion peaked after 6 h of contact and declined thereafter (Fig. S5A). The potential link between LAP and the secretion of proinflammatory cytokines/chemokines by astrocytes was then evaluated using siRNA targeting Ulk1, Rb1cc1, Rubcn/Rubicon and Atg5 genes. We found a significant down regulation of expression of Cxcl10, Ccl2 and Tnf/Tnf-α genes at 24 h after parasite contact in culture stimulated 6 h by PbA-iRBCs after treatment with siRNA targeting Rubcn/Rubicon and Atg5 genes when compared to control and Ulk1 and Rb1cc1 genes silenced astrocytes (Figure 9B). In concordance with protein expression data, the expression of Cxcl10, Ccl2, and Tnf/Tnf-α genes was elevated at 24 and/or 48 h post-PbA-iRBCs stimulation of astrocytes; this increase in expression was inhibited by 3-MA, an inhibitor of MAP1LC3/LC3 conversion (Figure 9C). By contrast, both Il1b/Il1β and Il6 gene expression increased after 3 h of co-culture, and 3-MA treatment increased Il6 but inhibited Il1b/Il1β gene expression (Fig. S5B). Except for Il6, a low increase expression of Il1b/Il1β, Tnf/Tnf-α, Cxcl10, and Ccl2 were observed in astrocytes stimulated 6 h with uninfected RBCs (Fig. S5C). These data were confirmed in astrocytes isolated from PbA infected B6 mice developing ECM as significant increase of Cxcl10 (P***<0.001), Ccl2 (P**<0.01), and Tnf/Tnf-α (P*<0.05) gene expression were detected compared to astrocytes from control mice (Figure 9D). Based on these observations, we concluded that the induction of pro-inflammatory cytokines and chemokines secretion by astrocytes in contact with PbA-iRBCs depended on LAP pathway involving the transfer of PbA-MVs.

Figure 9.

Role of PbA-MV˗induced autophagy in the proinflammatory response of astrocytes. Primary astrocyte cells were incubated with PbA-iRBCs at the ratio of 1:10 for 6 h and treated or not with 3-MA from the time of contact for up to 48 h. (A) Proinflammatory chemokines CXCL10, CCL2, and the cytokine TNF/TNF-α in the culture supernatants were measured by ELISA. (B) Astrocytes were transfected with Ulk1, Rb1cc1, Rubcn/rubicon, Atg5, or scramble siRNA. Afterward, PbA-iRBCs stimulation was performed for 6 h and cells were kept in new media to 24 h. Relative expression of Cxcl10, Ccl2 and Tnf/Tnf-α genes were quantified by RT-qPCR and compared to unstimulated astrocytes. (C) Expression of Cxcl10, Ccl2, and Tnf/Tnf-α genes in untreated and 3-MA treated astrocytes at different timepoints post-PbA-iRBCs contact. (D) ITGAM/CD11b⁻ SLC1A3/GLAST⁺ astrocytes were isolated from brain uninfected (control) and PbA infected B6 mice (n = 5 per group) and relative expression of Cxcl10, Ccl2, and Tnf/Tnf-α genes were quantified by RT-qPCR. Data are mean ± SEM for secreted proteins (n = 3 independent experiments) and median fold-change ± SEM of genes expressed (n = 5 independent experiments). Statistical analysis was done using one-way ANOVA for proteins and two-way ANOVA for gene expression, with Tukey’s multiple comparison test, with significance indicated at *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001. ns = not significant.

Discussion

Astrocytes are known to participate actively in brain inflammatory responses during ECM [8,15,32]. In this study, we used primary cultures and long-term live imaging to follow the interaction of astrocytes with PbA-iRBCs at the single-cell level. We showed that microvesicles (PbA-MVs) transferred in astrocytes upon contact with GFP-labeled PbA-iRBCs contain parasite material surrounded by PVM, as demonstrated by the presence of USI4, a known PVM marker [33]. This parasite material consisted of parasite DNA, as revealed by DAPI staining, and cytoplasmic constituents, evidenced by the presence of Pb18s genes and GFP. Astrocytes eliminated the parasite-derived MVs very efficiently, as shown by the simultaneous disappearance of GFP and DAPI signals and the decrease in Pb18s gene expression.

We then examined whether autophagy is involved in the degradation of PbA-MVs harbored within the astrocytes. Autophagy is a multi-step process that sequesters microbial pathogens in double-membrane vesicles called autophagosomes. These vesicles ultimately fuse with lysosomes to facilitate degradation of the pathogen cargo. By studying the sequence of molecular events that lead to the formation of the autophagosome, we showed that proteins involved in autophagy initiation and autophagosome biogenesis, such as BECN1, ATG16L1, and MAP1LC3/LC3, were increased after the transfer of PbA-MVs into astrocytes [18,34]. This was associated with the conversion of the LC3-I autophagy marker decorating the membrane of the autophagosome enclosing PbA-MVs to LC3-II. MAP1LC3/LC3 was coexpressed with UIS4, a marker of PbA PVM. We also observed a simultaneous decrease in the expression of SQSTM1/p62, a receptor and cargo protein known to regulate the packaging of substrates into the phagophore, and hence the delivery of the autophagosomal contents to the lysosome [18,35]. The perinuclear position of MAP1LC3/LC3⁺ autophagosomes, where lysosomes are enriched, and the decreased uptake intensity over time of LysoTracker Red provide evidence for lysosomal activity in PbA-MV-containing astrocytes. TEM analysis revealed that these autophagosomes were enclosed by a single membrane.

Our data supported the notion that a multistep autophagy process, which resulted in GFP-MAP1LC3/LC3 colocalization and which consisted of autophagosome nucleation, elongation, and maturation, took place during the transport of GFP⁺-PbA-MVs within the astrocytes. This process trafficked PbA-MVs from the iRBCs to the astrocyte plasma membrane and then further to lysosomes, where the parasite material underwent degradation. This mechanism seemed to be organized by recruitment of ATG proteins, and it started from 3 h via BECN1, followed by autophagosome vesiculation driven by ATG16L1 and ATG5, with LC3-I to LC3-II conversion and the involvement of SQSTM1/p62 receptor. These results were corroborated by the observation in cells treated with: i) 3-MA, an autophagy regulator that inhibits conversion of LC3-I to LC3-II, the formation of autophagosomes was prevented and PbA-MVs were blocked at the surface of the astrocytes, ii) BAF A₁, impairing the fusion of LAPosomes with lysosomes, PbA-MVs were accumulated in the cytoplasm of astrocytes. Conversely, treatment with rapamycin, which inhibits MTOR, enhanced autophagy flux. Altogether, these data suggest that the interaction of PbA erythrocytic stages with astrocytes, which are nonreplicating target cells, induces an autophagy-mediated host response. This response starts with the engulfment of PbA-MVs by the phagophore to generate a MAP1LC3/LC3⁺ autophagosome that is enclosed by a single membrane.

An autophagy-mediated host response has been previously described during hepatocyte invasion by PbA sporozoites wherein the successful development of hepatic stages is correlated with the gradual loss of canonical autophagy markers [24]. For their maturation and survival, growing parasites commandeer nonselective canonical host-cell autophagy as an additional source of nutrients. Sporozoites negatively modulate the host cell’s nonselective degradative machinery by impeding the fusion of autophagosomes with lysosomes and thereby allow for their own successful development into merozoites during the hepatic stages [24,36–40]. The fusion of autophagosomes with lysosomes is inhibited by the interaction of MAP1LC3/LC3 with UIS3 expressed in the PVM. Thus, UIS3 acts as an inhibitor of the autophagy machinery, thereby enabling the PbA sporozoite’s transformation into a trophozoite and its subsequent growth within the hepatocyte [24,41,42]. To avoid elimination, several pathogens have developed an ability to control and subvert the autophagy machinery by the circumvention of autolysosomal degradation [16]. This escape mechanism is often associated with disease severity [43]. For example, Toxoplasma gondii development in host cells has been shown to be dependent on the capacity of the parasite to prevent its killing by the host cell through activation of the EGFR (epidermal growth factor receptor)-AKT/protein kinase B (thymoma viral proto-oncogene) cascade that blocks the autophagosome-lysosome fusion. T. gondii can also use the nonselective host-cell autophagy machinery as an additional nutrient source [44,45].

We hypothesize that LAP is the autophagy process that mediated the formation of PbA-containing LAP-phagolysosomes and subsequent parasite degradation within the astrocytes [24,46]. Interestingly, LAP has also been observed during the development of PbA sporozoite in hepatocytes [24,30,39]. During LAP, autophagy proteins, including LC3-II, are recruited to the single-membrane LAPosome surface prior to its fusion with lysosomes, and this recruitment is independent of the activation of the ULK complex and associated downstream autophagy proteins [31,46,47]. Our hypothesis is strongly supported by our observation of the single membrane composition of PbA-MV containing LAPosomes decorated by MAP1LC3/LC3 and UIS4 as well as of the downregulation of Ulk1 and Calcoco2/Ndp52 gene expression in astrocytes upon PbA-iRBCs contact. CALCOCO2/NDP52 is an autophagy receptor-dependent ULK1 cascade gene, and both Ulk1 and Calcoco2/Ndp52 gene expression continued to be repressed when cell cultures were treated with rapamycin, an autophagy activator.

We show that BAF A₁ treatment protected mice from death caused by ECM without affecting the level of parasitemia. Proinflammatory factors such as CXCL10, CCL2, and TNF/TNF-α are involved in ECM [8,15]. Our finding that 3-MA treatment as well as silencing of RUBCN/rubicon and ATG5 by siRNA of astrocyte–PbA-iRBCs cultures suppressed the production of these proinflammatory factors strongly suggests that activation of the neuroinflammatory response of astrocytes that is observed during PbA-induced ECM could be dependent upon LAP. We therefore propose that parasite degradation through LAP by astrocytes could participate in the processing of PbA antigens and their presentation by MHC-I (major histocompatibility complex class I) molecules. Such increased expression of MHC-I molecules by activated astrocytes is known to occur during ECM [8]. LAP by astrocytes may thus trigger the recruitment of a selective repertoire of CD8⁺ T cells infiltrating the brain during CM [48–50]. However, the exact mechanism(s) responsible for triggering LAP remain to be determined. Preliminary results suggest that the TLR3 (toll like receptor) pathway activated in astrocytes during the PbA-MVs transfer could be involved in this process (our personal observation).

In summary, we provide the first evidence for the involvement of the unconventional autophagy pathway – LAP – that participate in the neuropathophysiological mechanisms of CM in driving the transfer/formation of PbA-MVs and the induction the proinflammatory response in astrocytes. This unconventional autophagy mechanism may thus contribute to antigen presentation and thereby promote the recruitment of pathological CD8⁺ T cells to the brain. Our data are strongly supported by the observation that an autophagy inhibitor that blocks the conversion of LC3-I to LC3-II strongly hampered PbA-MVs transfer into and subsequent secretion of proinflammatory mediators associated with ECM by astrocytes. This is consistent with the functional cooperation that is known to exist between glial cells and T lymphocytes in the brain, although the role of astrocytes as antigen-presenting cells during CM remains to be elucidated. Nevertheless, increased parasite MVs as well as astrocyte activation have been described in P. falciparum-infected patients developing severe disease [14,51,52]. In conclusion, our data indicate a role for LAP in the detrimental interplay of the malarial parasite with astrocytes that contribute to the development of CM. Therefore, LAP may constitute a potential therapeutic target to prevent the neuropathological mechanisms that precipitate CM.

Materials and methods

Mice

Female C57BL/6 mice, 8–10 weeks old (Janvier laboratories, C57BL/6JRjFEMELLESPF8), were maintained in the animal facility of the Pasteur Institute of Lille. Experiments were performed in agreement with the ethics of animal experimentation and approved by the French animal welfare committee “Ministère de l’Agriculture et de la Pêche” n°A 75485.

Infection of C57BL/6 mice with PbA-infected RBCs and treatment with bafilomycin A₁

Mice were infected intraperitoneally with 10⁶ red blood cells (RBCs) infected by the 1.49 L clone of PbA parasite expressing GFP (gift of Dr. D. Walliker, Institute of Genetics, Edinburgh, UK). Uninfected mice of the same age were used as controls. Mice were monitored twice daily, and parasitemia was assessed every 24 h by Giemsa-stained blood smears from the tail vein. All mice were euthanized when they started developing neurological symptoms characteristic of ECM.

For treatment with an autophagy inhibitor, mice received 0.1 mg/kg bafilomycin A₁ (BAF A₁; MedChemExpress, HY-100558) intraperitoneally daily, from day three to day five of the infection. Uninfected mice received 100 µL of phosphate-buffered saline (PBS; 1X; Gibco, 14,190,144). All mice were randomly allocated for treatments.

Isolation of P. berghei ANKA-infected red blood cells (PbA-iRBCs)

Severe combined immune deficiency mice (SCID mice; Institut Pasteur de Lille, SCID-8SEM-FEMELLES) were infected using RBCs parasitized by the transgenic GFP-expressing PbA clone (GFP-PbA) [53]. Blood was withdrawn at around 40–50% parasitemia in 50 µL of heparin (Sanofi-Synthélabo, 3,048,450), washed twice using PBS, and centrifuged at 300 g for 5 min at room temperature (RT). GFP-PbA-iRBCs were separated on 40% Percoll (Sigma, GE17-0891-01) by centrifugation at 700 g for 30 min. The washed iRBCs were resuspended in cDMEM (complete Dulbecco’s modified eagle medium; [Gibco, 11,885–084] containing 10% fetal bovine serum (FBS; [Dutscher, S181B-500])). About 95–100% purity was established by examining Giemsa-stained blood smears.

Primary culture of glial cells and enrichment of astrocytes

Primary glial cells were obtained from the brains of newborn (1–2 days old) C57BL/6 mice, gently dissociated under sterile conditions, and maintained in vitro as described previously [15].

Astrocytes were enriched in glial cell cultures using MACS Cell Separation anti-PE (phycoerythrin) mouse MicroBeads kit (Miltenyi Biotec, 130–048-801). Microglia were labeled by incubating the cells for 10 min at 4°C with PE-conjugated anti-ITGAM/CD11b-PE antibodies (eBioScience, 12–0112-81) in sterile PBS containing 2 mM of ethylenediaminetetraacetic acid and 0.5% FBS (PBS-EDTA). After a PBS-EDTA wash, cells were incubated with microbeads targeting the anti-ITGAM/CD11b-PE antibodies for 15 min at 4°C. Astrocytes were collected through negative sorting using a LS magnetic column (Miltenyi Biotec, 130–042-401) and cultured at a concentration of 1 × 10⁶ cells per dish in cDMEM with a weekly medium change. The enrichment of astrocytes was ascertained by flow cytometry (BD LSR Fortessa^TM, France) using anti-SLC1A3/GLAST (solute carrier family 1 (glial high affinity glutamate transporter), member 3)-allophycocyanine (APC) antibodies (Miltenyi Biotec, 130–095-814) to label astrocytes and anti-ITGAM/CD11b-PE antibodies to label microglia.

For stimulation, astrocytes were plated three days prior to the experimental time point at a concentration of 1 × 10⁶ cells/petri dish in cDMEM. Contact between GFP-PbA-iRBCs and astrocytes was performed at a ratio of 1:10. PbA-iRBCs were removed either 1) after 3 h (3 h timepoint) 2), or after 6 h (6 h timepoint), or 3) PbA-iRBCs were removed after 6 h, the medium was replaced, and the astrocytes were incubated further for 24 or 48 h (24 h and 48 h timepoints). The astrocytes were then washed twice with PBS and stored as cell pellets at −20°C.

For activation or inhibition of autophagy pathways, cells were either left untreated or treated with 5 mM rapamycin (Selleck Chemicals, S1039) or with 10 mM 3-methyladenine (3-MA; Selleck Chemicals, S2767) or with 10 nM of BAF A₁ (MedChemExpress, HY-100558) from the beginning of the 6-h stimulation.

siRNA silencing

Astrocytes were seeded at 5.10⁵ cells/well on a 12-well plate and 24 h later transfected with Ulk1, Rb1cc1, Rubcn/Rubicon or Atg5 siRNA or control siRNA at 25 nM using the HiPerfect transfection reagent (Qiagen, 301,704). Briefly, cells were incubated with the siRNA/HiPerfect mix for 6 h. Then, the medium was replaced by fresh medium and 48 h later, cells were stimulated with PbA-iRBCs or RBCs for 6 h. The medium was changed and quantifications of cytokine/chemokine production and autophagy gene expression were done 24 h later. Experiments were performed on five replicates. On-target plus SMARTpool siRNA (ULK1 22,241; RUBCN/rubicon 100,502,698; ATG5 11,793; RB1CC1 12,421) and control siRNA (D-001810-10-20) were purchased from Dharmacon (Horizon Discovery, UK).

Cell sorting for astrocytes isolation

Brain cells were isolated and purified using the trypsin-based Neural Tissue Dissociation kit (Miltenyi Biotec, 130–093-231) and 30% Percoll solution (Sigma, GE17-0891-01) before removing RBCs and live cells count using trypan blue by microscopy, as described previously [54]. The cell suspensions obtained from brain were stained with anti-ITGAM/CD11b (eBioScience, 12–0112-81) and anti-SLC1A3/GLAST (Miltenyi Biotec, 130–095-814) fluorescent antibodies and astrocytes ITGAM/CD11b⁻ SLC1A3/GLAST⁺ were sorted by FACSAria (Becton Dickinson) to be then analyzed by RT-qPCR, as described previously [54].

RNA extraction and quantification of expression

Astrocytes were isolated in vivo from the brains of CM⁺ or control mice and from primary cultures after contact with PbA-iRBCs or not (0 h). Total RNA was isolated using the Nucleospin RNA kit (Machery Nagel, 740,955.250). Complementary DNA (cDNA) was synthesized from 500 ng RNA using cDNA synthesis SuperScript VILO system (Invitrogen, 11,754,250) and a thermocycler C1000 (Bio-Rad) run: 10 min at 25°C, 60 min at 42°C, and 5 min at 85°C. About 2.5 ng of cDNA in the final 10 µL volume with primers (Table S2) and SYBER Green Master Mix (ThermoFisher Scientific, 4,385,612) were used for reverse transcription quantitative polymerase chain reaction (RT-qPCR). RT-qPCR was performed in Quantstudio^TM 12 K Flex Real-Time PCR system (ThermoFisher Scientific) and run for 10 min at 95°C, followed by 40 cycles of 15 sec at 95°C and 1 min at 60°C. Gene expressions were quantified by the ∆∆Ct and normalized to hypoxanthine phosphoribosyltransferase1 (Hprt1) expression for Becn1, Map1lc3/Lc3, Atg5, Atg16l1, Sqstm1/p62, Calcoco2/Ndp52, Rb1cc1, Rubcn/Rubicon, and Ulk1 genes for autophagy pathway; Pb18s; and Il1b/Il1β, Il6, Tnf/Tnf-α, Cxcl10, and Ccl2 genes for the cytokine/chemokine study. Data were analyzed with QUANTStudio software (ThermoFisher Scientific) and expressed as relative fold-change as compared to gene expression in unstimulated astrocytes.

Western blot analysis

Stimulated or unstimulated astrocytes were suspended in 200 µL lysis buffer (RIPA lysis buffer, Interchim, R0278-500 mL) with a cocktail of antiproteases (Roche, 04693159001) and resuspended in Laemmli buffer (Bio-Rad, 1,610,747). Proteins were separated on 10% or 15% acrylamide gels (Bio-Rad, 1,610,148) by SDS-PAGE and electro-transferred on 0.2-µm nitrocellulose membrane (ThermoFisher Scientific, 88,018). Membranes were blocked for 1 h with Tris Buffered Saline (TBS; 80 g/L sodium chloride [Fluka, 71,379] + 2 g/L potassium chloride [Sigma-Aldrich, P9541] + 30 g/L Trizma base [Sigma-Aldrich, T1503] diluted in demineralized water)-5% milk and then overnight at 4°C with 0.1% Tween 20® (Euromedex, 2001-B), before primary antibodies, anti-MAP1LC3/LC3 (1:2000; M186-3), anti-BECN1 (1:2000; PD017Y), anti-SQSTM1/p62 (1:2000; PM045Y), anti-ATG16L1 (1:2000; PM040Y), all from Medical & Biological Laboratories, and anti-ACTB/β-actin (1:3500; Cell Signaling Technology, 4970), were added. Following two washes with TBS-5% milk-0.1% Tween 20, horseradish peroxidase (HRP)-conjugated secondary antibodies-anti-rabbit IgG (Abcam, ab97051) or anti-mouse IgG (Abcam, ab6728), both at 1:10,000, were added for 1 h at RT. The blots were developed with Clarity^TM western ECL Substrate (Bio-Rad, 170–5061) and analyzed by the Molecular Imager ChemiDoc^TM XRS+ system (Bio-Rad) using Image Lab^TM software (Version 5.0, Bio-Rad). Proteins were quantified by densitometry (Version 1.52, ImageJ^TM software, National Institutes of Health) and intensities were normalized to that of ACTB/β-actin.

Confocal microscopy

Astrocytes were plated on glass slides and stimulated with PbA-iRBCs as described. They were then washed with PBS, fixed with 4% paraformaldehyde (Electron Microscopy Sciences, 15,713 R7G5) for 20 min, and washed again with PBS thrice. Cells were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich, T8787) in PBS for 5 min, washed, and incubated with rabbit anti-MAP1LC3/LC3 IgG (Medical & Biological Laboratories, PM036) or goat anti-UIS4 IgG (up-regulated in infective sporozoites gene 4) (MyBiosource Inc., MBS448030) in 10% FBS (Eurobio, 010056) for 1 h at RT. Cells were washed and incubated with anti-rabbit-IgG-Alexa Fluor 555 (Invitrogen, A21428) or anti-goat-IgG-Alexa Fluor 568 (Life Technologies, A11057) antibodies for 30 min at RT. After washing with PBS, cells were stained with rabbit anti-GFP-Alexa Fluor 488 (Invitrogen, A21311) antibodies for 1 h at RT before treating with DAPI (4ʹ,6-diamidino-2-phenylindole; 0.01 mg/mL; Invitrogen, D1306) for 15 min. Samples were observed with a Zeiss LSM880 confocal microscope (ZEISS microscopy GmbH) to obtain high-resolution images (voxel 0.045 × 0.045 × 0.80 μm³) using an Airyscan detector with a 63X oil immersion lens. Images were processed using ZEN software (Version2, ZEISS) for Airyscan processing using ImageJ for maximum intensity projection and rescaling of cropped images. Extended orthogonal views were produced using the Imaris software (Oxford Instruments) on a thickness of 5 μm centered on the object of interest. To quantify the presence of red-MAP1LC3/LC3⁺/GFP-PbA⁺ fluorescent vesicles in astrocytes, we performed an interactive image analysis by three counting steps to estimate: 1) total number of astrocytes; 2) the number of the GFP⁺-PbA-internalized MVs; 3) the number of red-MAP1LC3/LC3⁺MVs. Each stacking image was previously deconvoluted using Huygens Essential software (Scientific Volume Imaging) to minimize the contribution of artefactual external neighboring signal and to improve the accuracy of the fluorescent vesicles. Then, cell segmentation to generate an image of the intracellular MAP1LC3/LC3⁺ and GFP-PbA fluorescent spots and quantification were done using Imaris 9 single Full (BitPlane, Oxford Instruments).

Transmission electron microscopy (TEM)

Purified astrocytes were plated on coverslips in 35 mm glass bottom dishes (No 1.5, MatTek, P35G-1.5–10-C) and stimulated with GFP-PbA-iRBCs at 37°C for 6 h, then further incubated for 24 h. Cells were washed with PBS, fixed with 4% paraformaldehyde (Electron Microscopy Sciences, 15,713 R7G5) and labeled with DAPI (0.01 mg/mL; Invitrogen, D1306) and with rabbit anti-GFP-Alexa Fluor 488 (Invitrogen, A21311) antibodies to identify astrocytes containing GFP-microvesicles. Cells were fixed with 2.5% glutaraldehyde (Merck, 432VV351439), washed with 0.1 M phosphate buffer (Sigma-Aldrich, P5244-100 mL), and incubated with 0.1 M phosphate buffer containing 1% osmium (Alfa Aesar, 039998.06) and 1.5% potassium ferrocyanide (Merck, 60,279) for 1 h at RT. Fixed astrocytes were washed thrice with water and dehydrated using 50%, 70%, and 90% ethanol (Merck, 1.00983.1000), respectively, for 10 min, followed by two final dehydration steps at 100% ethanol. Inclusion of astrocytes was done using 50% ethanol and 50% of a premix agar resin (Agar Scientific, AGR1141) for 2 h at RT. The material was then put into 100% pure resin at RT for 2 h and finally incubated overnight at 60°C. The cell zones of interest were cut by a Diamond ultra 35 degree knife (Diatome) using Ultramicrotome UC6 (Leica Microsystems) with a step at 70 nm; the materials were collected on electron microscopy grids (Agar Scientific). Ultrathin sections were stained with uranyl acetate and lead citrate before observation with a JEOL1400 microscope operating at 80 kV and equipped with a RIO9 camera (Gatan).

Live microscopy

GFP-PbA-iRBCs were incubated with astrocytes at a ratio of 10:1 in glass bottom microwell dishes (MatTek, P35G-1.5–10-C). Four h after stimulation, the culture medium containing or not BAF A₁ at 10 nM was replaced with LysoTracker^TM Red DND-99 (Invitrogen, L7528) in cDMEM for 30 min. The cells were washed twice and placed in contact with GFP-PbA-iRBCs a second time to follow the steps in PbA-MVs transfer, with or without BAF A₁ treatment (MedChemExpress, HY-100558). The cells were imaged at 37°C and 5% CO₂, with snapshots captured every 20 min for 19 h. Images were acquired using a spinning disk Ti2-CSU-W1 (Nikon, Gataca Systems) with a 40× water immersion lens (Plan Fluar40×/1.30; Nikon). The emitted fluorescence was captured using sCMOS Prime 95B (Photometrics) camera. Movies were generated using ImageJ software.

Enzyme-linked immunosorbent assay (ELISA)

Mouse cytokines were measured in supernatants of unstimulated or PbA-iRBCs-stimulated astrocytes using ELISA kits for IL1B/IL1β (Biolegend, 430,901), IL6 (Biolegend, 431,301), CCL2 (Biolegend, 432,704), TNF/TNF-α (Biolegend, 432,601), and CXCL10 (R&D systems, DY466-05), according to the manufacturer’s instructions. Assays were performed in triplicates.

Statistical analysis

GraphPad Prism (Version 5, GraphPad software, Inc.) was used for statistical analyses, with data expressed as a mean or median ± standard error of the mean (SEM). The two groups were compared using the Student’s t-test. Analyses of variance (ANOVA), followed by Tukey’s multiple comparison or Bonferroni post-hoc tests, were used to compare kinetics groups and treatment groups. Comparison of survival were done using the Log-Rank (Mantel Cox) test. P values indicate statistical significance as *P < 0.05, **P < 0.01, ***P < 0.001, and ****P < 0.0001

Funding Statement

This work was supported by the LabEx PARAFRAP: LabEx PARAFRAP. [ANR-11-LABX-0024] and the GPF Brain and Infection of Institute Pasteur Paris. Institut Pasteur Paris [GPF Brain and Infection: NIMPALTOX].

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplementary material

Supplemental data for this article can be accessed here.