Astrocytes, the most abundant glia population in the central nervous system (CNS), contribute to several physiological functions oriented at maintaining the homeostasis of the CNS environment. Among these functions, astrocytes supply oligodendrocytes and neurons with substrates involved in energy metabolism, they control ionic and osmotic homeostasis, and they participate in synapse formation and neurotransmitter release. Astrocytes also modulate the local immune responses, and their cytoplasmic prolongations take part in the formation of the blood–brain barrier (BBB) . Given their broad spectrum of regulating activities, astrocyte dysfunction plays a pivotal role in the pathogenesis of several CNS disorders with widely diverging pathogenic mechanisms such as migraine, epilepsy, inflammatory demyelinating diseases (IDD), neurodegenerative disorders, and schizophrenia [1, 2].
Astrocytes are actively involved in the pathogenesis of multiple sclerosis (MS) and its murine model, experimental autoimmune encephalitis (EAE). MS is a chronic autoimmune inflammatory disease characterized by focal myelin loss and axonal degeneration [2, 3]. Effects of astrocytes on CNS immune-pathogenesis may include: (a) direct activation of T cells (as nonprofessional APCs) by presenting self-antigens in the context of MHC class II [3, 4], (b) secretion of chemokines such as CCL2, CCL5, and CXL10 that allows the attraction and migration of phagocytes and T lymphocytes to the CNS , (c) production of cytokines that help with the effector function of Th1, Th2, or Th17 populations [6, 7], and (d) production of the glial scar to build up a physical barrier around the demyelinated areas .
Other cells, such as macrophages and microglia, are found activated in demyelinating plaques in MS and they contribute to myelin phagocytosis and destruction, which are part of IDD pathogenesis [8, 9]. Rat macrophages and microglia considerably increase myelin phagocytic activity when cocultured with astrocytes or with astrocyte-conditioned medium, possibly due to a soluble factor secreted by these cells. As a consequence, there is a mechanism orchestrated by astrocytes that triggers the production of inflammatory factors in the presence of demyelination-derived products . It becomes relevant to explore what local factors in the CNS may play a role in generating astrocyte dysfunction. Myelin-derived proteins and their role in MS or EAE pathogenesis have been well documented . However, the influence of myelin-derived lipids on glia biology during demyelination has been studied mainly in microglia and macrophages [9, 12].
The myelin sheath is enriched in phosphatidylcholine, sphingomyelin (SM), cholesterol (CH), galactocerebrosides, and sulfated galactocerebrosides (sulfocerebrosides, SCB) [13-15]. Among these lipid components, SCB have received great attention because of their importance in the maturation of oligodendrocytes and the proper function of myelin . Disruption of sulfatide metabolism has been shown to result in defective myelin junctions. For example, mice deficient in UDP-galactose:ceramide (CER) galactosyltransferase (CGT), and cerebroside sulfotransferase (CST) activity, showed paranodal defects, which became more severe in CGT-null mutants .
SCB-containing liposomes avidly undergo endocytosis by macrophages, arguably through a receptor pathway . Also, SCBs trigger inflammatory (IL-6, TNF-α, and NO) release responses, mediated by NF-κB in rat microglia, and at least partly mediated through the binding to L-selectin (CD62L) . As astrocytes can be localized and proliferate in demyelinating plaques, they are likely to be in close contact with myelin or myelin derived lipids that could affect their biological behavior [8, 9]. Astrocytes are known to be involved in lipid homeostatic routes in the CNS such as the production of ApoE [18-21]. ApoE has been shown to transport sulfatides from the myelin sheath into neurons, where ApoE/SCB undergo endocytosis mediated by the low density lipoprotein receptors (LDL-R) [18, 19]. Others reported that murine adult astrocytes can undergo CH endocytosis using scavenger receptors and CD36 . Here, we study the uptake of liposomes of several myelin-related compositions by human astrocytes to explore whether myelin lipids influence liposome uptake rates by astrocytes as indication of the presence of selective pathways for lipid identification. The influence of these liposomes in cell viability, human leukocyte antigen (HLA) molecules expression, and chemokine secretion was also determined with the aim of exploring whether the uptake of myelin-derived lipids can induce an inflammatory response in astrocytes.
Materials and Methods
Lipids and Reagents
1,2-Dioleoyl-sn-glycero-3-phosphatidylcholine (DOPC), 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine (POPE), porcine brain SM, N-octadecanoyl-D-erythro-sphingosine (porcine brain CER), cerebroside (CB), sulfocerebroside, and CH were purchased from Avanti Polar Lipids (Birmingham, AL) and used without further purification. N-4-Nitrobenz-2-oxa-1,3-diazolyl phosphatydilethanolamine (NBD-PE) was purchased from Invitrogen (Carlsbad, CA). Vesicles were named DOPC, POPE, CER, SCB, and CB, and their compositions are shown in Table 1. Recombinant human IFNγ was obtained from Miltenyi Biotec (Bergisch Gladbach, Germany), fluorescent spheres from Spherotec Inc. (Lake Forest, IL), and propidium iodide (PI) from (DakoCytomation, Glostrup, Denmark). Antibodies for flow cytometry and cytometry beads arrays (CBA) were purchased from BD Pharmingen (San Diego, CA); and BD Bioscience (San Jose, CA). Other chemicals and reagents for culture medium, including cytochalasin D, were acquired from Sigma-Aldrich (St. Louis, MO), Eurobio (Les Ulys, France), or Gibco (Auckland, New Zealand).
Table 1. Lipid composition of liposomes
Lipid composition (%)
Lipid compositions used to prepare liposome samples. Liposomes are also labeled with 2 mol % NBD-PE. Values indicate proportions between different components.
Abbreviations indicate the myelin lipid component that is being tested with this composition. CB, cerebrosides; CER, ceramide; CH, cholesterol; DOPC, 1,2-dioleoyl-3-sn-phosphatidylcholine; POPE, 1-palmitoyl-2-oleoyl-sn-glycero-3-phosphoethanolamine; SM, sphingomyelin; SCB, sulfocerebrosides.
Lipids with 2 mol % NBD-PE were mixed in chloroform in the proportions shown in Table 1, and dried under a stream of nitrogen gas. The lipid films were then cooled to −25°C for 2 h and lyophilized for at least 6 h before use. The lipid samples were then dispersed in vesicle buffer (100 mM NaCl, 10 mM HEPES, 1 mM EDTA, and 1 mM CaCl2 at 7.4 pH) at a 3 mM concentration; and 1.1 ml of the sample was mixed by vortex for 1 min, followed by 15 min of rest. This process was repeated five times. The sample underwent sonication in a VCX 750 tip sonicator (Sonics) with 10 s/10 s pulses at minimum (4–6 Watts) power output using a 3-mm tapered tip. The average vesicle size was determined in a particle size analyzer (SALD 7001, Shimadzu), and they measured 58 ± 8 nm.
Astrocyte Culture and Vesicle Uptake Kinetics Assay
The human astrocytes cell line used in all experiments was CRL-1718 from ATCC (Manassas, VA). Cells were maintained in RPMI-1640 (Sigma, St. Louis, MO) medium with 10% of foetal calf serum (FCS) (Eurobio, Les Ulys, France), glucose at 4.5 g/l (Sigma), 10 mM HEPES, 2 mM L-glutamine, 1 mM sodium pyruvate (Gibco, Auckland, New Zealand), and cultivated in T25 flasks at 37°C and 5% CO2. Following monolayer formation, cells were detached with 0.25% trypsin-EDTA (Gibco) for 3 min at 37°C and then recovered in RPMI-1640 medium with 5% of FCS. Cell viability was assayed by trypan blue exclusion staining. Detached cells (1 × 105/ml) were cocultivated with 0.2 mM of NBD-PE labeled (2 mol %) fluorescence vesicles, and readings were carried out with an ISS PC1 Spectrofluorimeter (Champaign, IL), with excitation and emission wavelengths was set at 470 and 540 nm, respectively. Measurements for vesicle uptake were performed at 30 min, 1, 2, 3, and 4 h after incubation. At each time point, cells were washed with 10 ml of 1× PBS and resuspended in vesicle buffer. Unlabeled cells were used as a negative control. For flow cytometry, at least 5–10 × 103 cells were acquired in FACs Canto II (BD Bioscience, San Jose, CA). For some experiments, a U937 promonocyte cell line (CRL-1593, ATCC) was used and maintained in RPMI-1640 with 5% FCS. Flow cytometry data were analyzed with the FACs DIVA software (BD Bioscience). Cells were characterized by their relative size (forward scatter, FSC) and relative granularity (side scatter, SCC). NBD-PE fluorescence was determined using the 530 nm wavelength detector in FACs Canto II, the flow cytometry gating strategy is shown in the Supporting Information Figure 1. In some inhibition experiments, astrocytes were preincubated for 3 h at 4°C or 30 min with 40 μM of cytochalasin D before adding the SCB liposomes. Cells were washed with PBS 1× (cold for cells incubated at 4°C), and after incubation of 1 or 2 h, the fluorescence was measured by flow cytometry as described above. For these mentioned experiments, controls were done at room temperature without cytochalasin D.
Astrocytes were cultured in glass bottom (35 mm) microwell dishes (MatTek Corporation, Ashland, MA) in the same conditions as described above. Pictures were taken at different time points after incubation with lipid vesicles and unfixed cultures in a Nikon Eclipse TI microscope at 60× (Nikon Instruments, Melville, NY).
Astrocyte Cultures for Functional Assays
Astrocytes were cultured in 48-well plates (Falcon, Franklin Lakes, NJ) at 1 × 105 cells/ml and incubated for 48 h in triplicates, alone or with 0.2 mM lipid vesicles. As a positive control, astrocytes were cultured in the presence of 1µg/ml IFNγ (Miltenyi Biotec, Bergisch Gladbach, Germany). At least three independent experiments were performed for the functional assays such as cell viability, cell proliferation, MHC expression, and cytokines/chemokines measurements.
Bead Uptake by Human Astrocyte and Promonocyte Cell Lines
Nonstimulated U937 cells and IFNγ-activated astrocytes were used for uptake of fluorescent bead assays using SPHERO Fluorescent Yellow Particles of 0.049 µm (Spherotec, Lake Forest, IL). Beads (1 µl) were added to each culture and incubated for 24 h at 37°C. Following the incubation, the cells were washed and analyzed in a flow cytometer to determine the bead capturing cell percentage.
Cell Counting and Viability
Cells in 48-well plates were detached following 48 h of culture after lipid vesicle incubation as described above. After washing with RPMI-1640 containing 5% FCS, cells were counted in a Neubauer chamber in the presence of trypan blue to obtain the cell number per milliliter. The percentage of viability was determined by PI staining (DakoCytomation, Glostrup, Denmark) by flow cytometry.
HLA Class I and II Molecule Expression
Following 48 h of incubation with vesicles, cells were tested for HLA class I and class II expression by flow cytometry. FITC-labeled anti-HLA-ABC antibody (clone G46-2.6, BD Pharmingen, San Diego, CA) was used for class I, and PE-Cy7-labeled anti-HLA-DR (clone L243, BD Pharmingen) for HLA class II detection. Dead cells were excluded by PI staining. The percentage of expression and mean fluorescence intensity (MFI) for each fluorochrome was determined by flow cytometry. The flow cytometry gating strategy is shown in the Supporting Information Figure 1.
Chemokine and Cytokine Levels
Sterile supernatants obtained after 48 h of incubation with lipid vesicles were fractioned and stored at −80°C until used. Chemokines and cytokines were measured in 50 µl of supernatant in each replica using the BD Cytometric Bead Array, Human Chemokine Kit-I (CXCL10/IP-10, CCL2/MIP-1, CXCL9/MIG, CCL5/RANTES, and CXCL8/IL-8) and Human Inflammatory Cytokine CBA Kit (IL-8, IL-1b, IL-6, IL-10, TNF, IL-12p70) (BD Bioscience, San Diego, CA). The CBA assay was read in the FACs Canto II cytometer and analyzed with the FCAP Array program 1.01 version (BD Bioscience).
Results are shown using descriptive statistics (mean and standard deviation). For group comparisons (molecule expression, cell count, and viability), nonparametric statistics were used, Kruskal–Wallis all pairwise comparison test, considering a P < 0.05 value as significant.
We explored whether human astrocytes are capable of internalizing protein-free unilamellar vesicles (liposomes) of various lipid compositions, and whether this process induces changes in the biological function of the astrocytes or triggers a proinflammatory response. Cells were exposed to NBD-PE-labeled liposomes for different time periods, and liposome uptake was detected using fluorescence spectrofluorometry (Fig. 1a) and flow cytometry (Fig. 1b). Different lipid composition analogues to human myelin were tested (Table 1); where the molar fraction of each sphingolipid in the liposomes reflects the reported composition of the human myelin sheath . Liposomes composed of DOPC/SM/CH (46:29:25) were used as a control. This composition reflects the most common lipid components of the external leaflet in human plasma membrane . Phosphatidylethanolamine (POPE)-containing liposomes [DOPC/POPE/SM/CH (35:30:15:20)] were used as a positive control, due to the propensity of POPE to promote membrane fusion as a result of its inverted conical shape . The presence of POPE is, therefore, expected to enhance liposome uptake.
After exposure to liposomes, an initial rise in the uptake level for all vesicle compositions was observed, reaching saturation within 1 h (Figs. 1a and 1b). Clear differences in the saturation levels were detected for the different lipid compositions. Liposomes containing CER and cerebrosides (CB) reached saturation levels similar to the control composition vesicles (DOPC). In contrast, SCB containing liposomes displayed a higher uptake level than those observed for the control (DOPC), also surpassing the uptake observed for POPE-containing vesicles (Figs. 1a and 1b). Spectrofluorometry and flow cytometry show the same uptake trend for the various lipid compositions.
Cells were visualized by fluorescence microscopy at 6 and 24 h after exposure to NBD-labeled DOPC liposomes to determine how liposomes are internalized by human astrocytes. After 6 h, liposomes are found spread throughout the cytoplasm including the cell membrane prolongations (Figs. 2b and 2c). After 24 h, liposomes are found localized around the cell nucleus (Figs. 2e and 2f).
In some experimental controls, human monocytes (U937 cell line) exposed to DOPC liposomes were used. These presented faster saturation (30 min) and higher uptake than human astrocytes (Supporting Information Fig. 2). This is expected since monocytes are proficient phagocytic cells. To test the specificity of the liposome uptake, astrocytes, and U937 monocytes were exposed to 49-nm fluorescent polystyrene beads. These beads are in the same size range as the sonicated liposomes used in this study (vesicle average size 58 nm). Monocytes presented high levels of bead uptake, as measured by flow cytometry (Fig. 3a), whereas astrocytes did not show significant bead uptake (Fig. 3b), even after activation with IFNγ (Fig. 3c). As beads were not taken up by the astrocytes, these results suggest that lipid composition can specifically regulate internalization of the liposomes by astrocytes.
Astrocyte cultures, exposed to liposomes for all the compositions described, did not show any difference in cell proliferation or mortality after 48 h of incubation (data not shown). Expression of HLA class I (HLA-ABC) and class II (HLA-DR) molecules was assessed in astrocytes cultured in the presence of liposomes. The percentage of cells expressing HLA class I molecules was always above 95%, remaining constant for all liposome compositions, and for control astrocytes cultured in the presence of IFNγ (data not shown). However, when HLA class I mean fluorescent intensity (MFI) was determined, astrocytes exposed to IFNγ showed as expected increased fluorescent intensity (P = 0.0227). When comparing untreated astrocytes to liposome-exposed astrocytes, no increase in HLA class I MFI expression levels were detected (Fig. 4a). Likewise, the percentage of cells expressing HLA class II molecules increased notably in the presence of IFNγ (P = 0.0147), whereas HLA class II molecule expression in astrocytes remained at basal levels (similar to untreated cells) for all liposome compositions (Fig. 4b).
Astrocytes are capable of producing several chemokines and cytokines, some of them important for the recruitment of leukocytes during the CNS inflammation. Human cytokines (IL-1β, IL-10, IL-12p70, and TNFα) were not detected in the supernatant culture after 48 hrs of incubation with liposomes. Only IL-6 was observed in the IFNγ treated cultures (data not shown). Astrocytes alone or in the presence of liposomes produce similar levels of CXCL8 (IL-8), CCL2 (MCP-1), and CXCL10 (IP-10), but did not produce CCL5 (RANTES) or CXCL9 (MIG). Stimulation of astrocytes with IFNγ induced the secretion of CCL5 and CXCL9. Furthermore, stimulation induces an up regulation of CCL2 (from 5,100.4 to 8,007.4 µg/ml) and CXCL10 (from 3,472.4 to 5,304.4 µg/ml). Despite the fact that astrocytes produce chemokines, liposome uptake did not induce changes in their secretion (Fig. 5).
Possible astrocyte uptake routes appear to be modulated by the presence of SCB in the liposomes. To further test the specificity of this uptake route, human astrocytes where exposed to different concentrations of SCB before treating the cultures with SCB liposomes (Fig. 6). The results show a dose dependent reduction in vesicle uptake as a function of exogenous addition of SCB dissolved in DMSO, reaching saturation around 10 μM. Finally, two experiments were conducted as an initial exploration to determine the mechanisms of liposome internalization. Cultures were either incubated at low temperature (4°C) to block endocytosis or preincubated with 40 µM of cytochalasin D to inhibit actin-mediated endocytosis. Results are summarized in Table 2, and are expressed as percentage of change (taking control experiments as 100%). Only a slight increase in liposome uptake was observed in cultures treated with cytochalasin D. Thus, blocking the actin-mediated endocytosis pathway does not appear to significantly alter SCB lipid uptake. In experiments in cells incubated at 4°C, there was a decrease in the MFI (2,601.5 a.u. ± 194.3 at 1 h; 2,732 ± 323.9 at 2 h) in SCB liposome uptake compared to 25°C (4,876.5 ± 347.7 at 1 h; 6,063.3 ± 854.1 at 2 h). On average, the percentage of inhibition was near 50% in both time points (Table 2). As endocytosis is expected to be fully inhibited at 4°C, this 50% SCB liposome uptake was unusual. Further investigation revealed that this fluorescence signal detected by flow cytometry was from liposomes bound to the cell membrane surface, that did not undergo endocytosis. Epifluorescence microscopy images of astrocytes exposed to SCB liposomes at 4°C show that liposomes are bound around the cell perimeter, (Figs. 7b and 7c) delimiting the cell body and not inside the cytoplasm. In contrast, when treated at room temperature SBC liposomes are found spread within the astrocyte's cytoplasm and their prolongations (Figs. 7e and 7f).
Table 2. Percentage of changes in SCB liposomes uptake in astrocytes incubated at 4°C or with an actin-inhibitor
aAverage of two independent experiments done in duplicates.
+29.8 ± 9.1
−46.6 ± 3.3
+26.4 ± 8.4
−54.6 ± 5.6
Human astrocytes, the most abundant cells in the CNS, present mechanisms for internalization of extracellular proteins  and lipids . Astrocytes are involved in several important biological functions in the CNS, which include control of neurotransmitters, protection of neurons, local immune response, and control of the BBB, among others. Perturbing the function of astrocytes contributes to CNS physiopathology during infectious, vascular, degenerative, and autoimmune diseases [1, 3]. During inflammatory and demyelinating CNS diseases, astrocytes produce several chemokines and cytokines . Astrocytes can also be nonconventional antigen-presenting cells, although they express HLA class II molecules in low percentages in vivo . The production of cytokines and HLA molecules by astrocytes has been described mainly in models where protein-derived myelin antigens induce activation of the immune response [4, 11]. Astrocyte response to lipid debris, derived from the demyelination process, has not received much attention. However, electron microscopy studies have shown the presence of membrane-bound vesicles filled with myelin in close relationship with astrocytes during demyelination . This led us to assess if the contact of myelin-derived lipids with these cells can induce changes that favor a proinflammatory response. Here, it is shown that astrocytes can internalize liposomes. The process was found to be lipid specific since fluorescent beads where not internalized. Additionally, the rate of liposome internalization is composition dependent, with a particular myelin-specific lipid, sulfocerebroside, augmenting the uptake process. Enrichment of liposomes around the cell nuclei suggests that internalization concurs through an intracellular mechanism of active transport, although blocking the actin-dependent endocytosis pathway does not decrease liposome internalization. The liposome lipid composition used did not induce secretion of proinflammatory chemokines and cytokines. Also, expression of HLA class I and class II HLA molecules, which are induced by IFNγ in the astrocyte cell line, remain constant when exposed to liposomes . These results suggest that lipid-derived debris produced during demyelination do not appear to participate in inducing CNS proinflammatory responses through astrocytes' activaction. This is opposite to microglia, which can be activated by sulfatides, inducing the secretion of TNF-α, IL-6, IL-8, and MCP-1 . Interestingly, myelin foamy macrophages, found in brain tissue of patients with MS, express higher levels of anti-inflammatory cytokines such as IL-4, IL-10, and TGF-β  which would correlate with the capacity of myelin-derived debris to actively infiltrate macrophages and local microglia.
As human astrocytes do not respond to lipids in the same manner as CNS microglia/macrophages, what role would this lipid-specific uptake mechanism play in the CNS physiology? Although their role in pathogenesis is not known, it has been documented that sulfatides and CER are found in reduced levels in the CNS of individuals with Alzheimer's disease [18, 19], indicating their essential participation in the proper function of the CNS. A tight control of lipid levels in the different CNS cells is therefore needed. Current knowledge indicates that astrocytes participate in CNS lipid homeostasis [19, 29-31]. For instance, astrocytes take part in sulfatide glucocerebroside transport by synthesizing the apolipoprotein (Apo) E [19, 21]. Apo E sequesters sulfatides from the myelin sheath, delivering the lipids to the neighboring neurons [14, 19, 32]. Our results point to an additional metabolic pathway, involving uptake of SCB by astrocytes. This suggests the presence of a metabolic route for processing sulfatides that may play a role in CNS lipid homeostasis. All lipid compositions tested here were internalized by astrocytes. However, these cells more avidly internalize sulfocerebroside-containing liposomes. Two simple approaches were used to assess liposome uptake: blocking endocytosis at 4°C and inhibiting actin-dependent endocytosis by cytochalasin D. Low temperatures reduced SCB liposome incorporation into astrocytes by about 50% when measured by flow cytometry, but fluorescence microscopy images showed liposomes bound to the outer leaflet of the plasma membrane without being internalized . This points to the presence of an active transport system for liposome internalization. However, this endocytic pathway does not appear to involve actin polymerization.
Regarding lipid trafficking, sulfatides are mainly synthesized by oligodendrocytes, with residual amounts produced by neurons. However, in rat astrocytes, electron microscopy images show the presence of sulfatides in intracellular compartments [29, 31], where it is not clear if sulfatides are being synthesized by the astrocytes or recycled from the CNS lipid environment. In the case of neurons, the complexes of Apo E/SCB can be taken up and internalized through the low density lipoprotein receptors (LDL-R) [19, 36, 37]. ApoE knockout mice displayed altered SCB content in brain tissue without presenting noticeable alterations in other CNS lipids . Upon internalization by neurons, degradation of SCB takes place in lysosomes where arylsulfatase A (ASA) hydrolyzes the sulphate group. In ASA deficiency, there is an accumulation of SCB in CNS cells causing a severe lysosome storage disease-named metachromatic leukodystrophy (MLD) [18, 38].
Human astrocytes express the LDL receptor under normal conditions [37, 39] and its expression can be upregulated by IFNγ , a cytokine produced by T cells that infiltrates the CNS during pathological conditions . Also, astrocytes have ASA and sulfatides accumulated during ASA deficiency . Based on the literature and the results presented here, it is plausible to suggest that astrocytes present a sulfocerebroside-specific uptake mechanism that may modulate the trafficking of sulfocerebroside as a part of the lipid metabolism in the CNS. Similar outsourcing has been proposed for other lipids such as CH [20, 36]. Of notice, soluble CER has been found to induce apoptosis in neurons , neuronal tumor-derived cell lines , and glia such as oligodendrocytes, but not in astrocytes . Here, astrocytes did not undergo apoptosis when exposed to liposomes containing CER and SCB. Although CER do not induce up regulation of lipid uptake in astrocytes, the absence of an apoptotic response reinforces the notion of the involvement of astrocytes in CNS lipid homeostasis.
The authors do not report any conflict of interest. We want to thank Steve Trier Ph.D. Department of Physics, Universidad de los Andes for reading the manuscript.