Abstract
Activated microglia are thought to undergo apoptosis as a self-regulatory mechanism. To better understand molecular mechanisms of the microglial apoptosis, apoptosis-resistant variants of microglial cells were selected and characterized. The expression of lipocalin 2 (lcn2) was significantly down-regulated in the microglial cells that were resistant to NO-induced apoptosis. lcn2 expression was increased by inflammatory stimuli in microglia. The stable expression of lcn2 as well as the addition of rLCN2 protein augmented the sensitivity of microglia to the NO-induced apoptosis, while knockdown of lcn2 expression using short hairpin RNA attenuated the cell death. Microglial cells with increased lcn2 expression were more sensitive to other cytotoxic agents as well. Thus, inflammatory activation of microglia may lead to up-regulation of lcn2 expression, which sensitizes microglia to the self-regulatory apoptosis. Additionally, the stable expression of lcn2 in BV-2 microglia cells induced a morphological change of the cells into the round shape with a loss of processes. Treatment of primary microglia cultures with the rLCN2 protein also induced the deramification of microglia. The deramification of microglia was closely related with the apoptosis-prone phenotype, because other deramification-inducing agents such as cAMP-elevating agent forskolin, ATP, and calcium ionophore also rendered microglia more sensitive to cell death. Taken together, our results suggest that activated microglia may secrete LCN2 protein, which act in an autocrine manner to sensitize microglia to the self-regulatory apoptosis and to endow microglia with an amoeboid form, a canonical morphology of activated microglia in vivo.
Microglia are the CNS-resident immune cells (1, 2, 3). These phagocytic cells function as the first line of defense in the CNS. When stimulated, ramified resting microglia transform into the amoeboid form (4) and they actively participate in the CNS immune and inflammatory responses (5, 6). Although the morphological transformation from the ramified microglia into the amoeboid microglia is observed under a wide variety of in vitro as well as in vivo conditions (7, 8, 9), the precise mechanism of the changes in the microglial morphology remains largely unknown. Activated microglia migrate to area of injured nervous tissue and they engulf and destroy microbes and cellular debris (10). The role of microglia as inflammatory cells is not always beneficial. Uncontrolled microglial activation and the subsequent excessive neuroinflammation are now believed to contribute to a variety of the CNS pathologies including neurodegenerative diseases (11, 12, 13, 14). Thus, inflammatory activation of microglia has to be strictly regulated, and the apoptotic elimination of activated microglia is thought to be one mechanism of the microglial self-regulation (15, 16, 17, 18, 19). Evidence has been provided that inflammatory cells in the CNS may undergo apoptosis upon activation in a manner similar to activation-induced cell death of lymphocytes (16, 20, 21). However, little is known about the molecular mechanisms underlying the termination of neuroinflammation and the autoregulatory apoptosis of microglia.
Lipocalin 2 (LCN2)3 is a member of lipocalin family, which binds or transports lipid and other hydrophobic molecules (22, 23). LCN2 is also known as 24p3 (24), 24-kDa superinducible protein (SIP24) (25), and neutrophil gelatinase-associated lipocalin (NGAL; a human homolog of lcn2) (26, 27). LCN2 has diverse functions. In vitro studies have shown that LCN2 is important for both cellular apoptosis and survival (28, 29, 30, 31, 32). It also plays a central role in the induction of cellular differentiation in the kidney during embryogenesis (33) and protects the kidney from ischemic injury (34, 35). In various forms of gastrointestinal injury, LCN2 facilitates mucosal regeneration by promoting cell migration (36). In vivo studies based on lcn2-deficient mice, however, argued against the role of LCN2 in the renal protection (37). Moreover, lcn2-deficient mice exhibited an increased susceptibility to bacterial infections because of the failure of iron sequestration, indicating a critical role of lcn2 in protection against bacterial infection (38). Previous works have also suggested that lcn2 protects neutrophil gelatinase from degradation (so, the name NGAL) (39), and it may function as an acute phase protein (40). Recently, two cellular receptors for lcn2 have been identified. Megalin, a member of the low-density lipoprotein receptor family, has been shown to bind human lcn2 and to mediate its cellular uptake (41). Brain type organic cation transporter is another cell surface receptor for mouse lcn2, which has been shown to selectively mediate apoptosis (29). Despite the receptor identification, the precise role of lcn2 in cell survival and death has yet to be determined.
In the present study, we used a transcriptomic approach combined with the phenotypic selection of apoptosis-resistant microglial cells to better understand the mechanisms of microglial apoptosis. One of the genes that was significantly down-regulated in the apoptosis-resistant microglial cells was lcn2. Subsequent studies including the forced expression and knockdown of gene expression revealed that lcn2 is critical for apoptosis sensitization as well as for the amoeboid transformation of activated microglia.
Materials and Methods
Reagents and cells
LPS from Escherichia coli 0111:B4 prepared by phenolic extraction and gel filtration chromatography, S-nitroso-N-acetylpenicillamine (SNAP), sodium nitroprusside (SNP), etoposide, cisplatin, hydrogen peroxide, PMA, deferoxamine mesylate (DFO), thrombin, ATP, A2318742) which exhibits phenotypic and functional properties comparable to those of primary microglial cells (43) were grown in DMEM containing 5% FBS, 2 mM glutamine, and penicillin-streptomycin (Invitrogen Life Technologies). Mouse primary microglial cultures were prepared by mild trypsinization as previously described with minor modifications (44). In brief, forebrains of newborn ICR mice were chopped and dissociated by mechanical disruption using a nylon mesh. The cells were seeded into poly-l-lysine-coated flasks. After in vitro culture for 10–14 days, microglial cells were isolated from mixed glial cultures by mild trypsinization. Mixed glial cultures were incubated with a trypsin solution (0.25% trypsin, 1 mM EDTA in HBSS) diluted 1/4 in PBS containing 1 mM CaCl2 for 30–60 min. This resulted in the detachment of an upper layer of astrocytes in one piece, whereas microglia remained attached to the bottom of the culture flask. The detached layer of astrocytes was aspirated and the remaining microglia were used for experiments. The purity of microglial cultures was >95% as determined by isolectin B4 staining. Animals used in the current research have been acquired and cared for in accordance with the guidelines published in the National Institutes of Health Guide for the Care and Use of Laboratory Animals.
Microarray analysis
For the gene expression profiling studies, DNA microarray analysis was conducted by GenoCheck using a Platinum Biochip mouse 7.4 k cDNA chip containing 7636 cDNA spots. Data analysis was done using GenePix pro 4.1 software. Log-transformed spot intensity was plotted (M vs A scatter plot), normalized, and further analyzed as previously described (45). A complete listing of the genes on the microarray and detailed experimental protocols is available at the GenoCheck web site (www.genocheck.com).
Assessment of cytotoxicity by MTT assay or TUNEL staining
For MTT assay, cells (5 × 104 cells in 200 μl/well) were seeded in 96-well plates and treated with various stimuli for the indicated time periods. After treatment, the medium was removed and MTT (0.5 mg/ml) was added, followed by incubation at 37°C for 2 h in CO2 incubator. After insoluble crystals were completely dissolved in DMSO, absorbance at 570 nm was measured using a microplate reader (Anthos Labtec Instruments). Apoptosis of microglia was scored by TUNEL assay using a commercially available kit according to the manufacturer’s protocol (In Situ Cell Death Detection kit, POD; Roche Applied Science). The percentage of apoptotic cells was quantitated by counting TUNEL-positive cells in 10 random microscope fields.
Evaluation of mitochondrial membrane potential
Mitochondrial potential was measured using a lipophilic cationic probe 5,5′,6,6′-tetrachloro-1,1′,3,3′-tetraethylbenzimidazolcarbocyanine iodide (JC-1; Molecular Probes). After incubating the cells with 10 μg/ml JC-1 at 37°C for 10 min, emission at 525 and 590 nm was measured using a fluorescent microplate reader. The ratio of A590 (red) to A525 (green) was calculated as an indicator of mitochondrial potential. A decrease in membrane potential leads to a decrease in the ratio of red:green fluorescence.
Morphological analyses of microglial deramification and apoptosis
Morphological analysis of microglia was performed using phase contrast and fluorescence microscope. Isolectin B4 staining was done as previously described (46). Microscopic images were processed using MetaMorph Imaging System (Molecular Devices). Deramification of microglia was quantitated as previously described with a slight modification (47). Ramified cells were defined as having at least two processes, among which one process was longer than one cell body diameter. Nonramified cells were those that did not fulfill these criteria. The percentage of ramified cells was determined from a minimum of five randomly chosen fields containing at least 250 cells. Morphological changes in the nuclear chromatin of cells undergoing apoptosis were detected by staining with 2.5 μg/ml Hoechst 33342 fluorochrome (Molecular Probes), followed by examination on a fluorescence microscope (Olympus BX50).
Flow cytometric analysis of apoptosis
Microglia were detached with trypsin-EDTA and washed twice with cold PBS. The cells were then resuspended in 250 μl of binding buffer (10 mM HEPES, 140 mM NaCl, and 2.5 mM CaCl2 (pH 7.4)) and incubated with 3 μl of FITC-conjugated annexin V (Molecular Probes) according to the manufacturer’s specifications. Afterward, cells were gently vortexed and incubated for 15 min at room temperature in the dark. Propidium iodide (20 μg/ml) was then added and flow cytometry was performed within 1 h using FACSAria (BD Biosciences).
RT-PCR
Total RNA was extracted from BV-2 cells or primary microglial cells by TRIzol reagent (Invitrogen Life Technologies) according to the manufacturer’s protocol. Reverse transcription was conducted using Superscript (Invitrogen Life Technologies) and oligo(dT) primer. PCR amplification using specific primer sets was conducted at 55°C annealing temperature for 30–40 cycles. Nucleotide sequences of the primers were based on published cDNA sequences (Table I⇓).
DNA sequences of the primers used for RT-PCR
Western blot analysis
g (5 min), 15,600 × g (10 min), to remove nonadherent cells and debris. Samples were then precipitated with a trichloroacetic acid and acetone mixture (10% trichloroacetic acid and 10 mM DTT in acetone) at −20°C overnight. Precipitated proteins were subjected to the SDS-PAGE and ECL detection.
Stable transfection of lcn2 cDNA or short hairpin RNA (shRNA)
BV-2 cells in 6-well plates were transfected with 2 μg of mouse lcn2 or lcn2-specific shRNA constructs using Lipofectamine reagent (Invitrogen Life Technologies). Mammalian expression construct of mouse lcn2 cloned into pcDNA3 was a gift from Dr. J. P. Kehrer (University of Texas, Austin, TX) (32), and lcn2-specific shRNA construct was a gift from Dr. L. G. Cantley (Yale University, New Haven, CT) (48). The shRNA corresponding to the nucleotides 192–213 of lcn2 cDNA followed by a 9-base loop and the inverted repeat was cloned into pSuppressor Retro vector (Imgenex). An empty pcDNA3 vector and a vector encoding the scrambled RNA were used as a control for the stable expression of lcn2 and lcn2-specific shRNA, respectively. Stable transfectants were selected in the presence of G418 (400 μg/ml) at 2 days after the transfection. Up- or down-regulation of LCN2 protein in the stable transfectants was confirmed by Western blot analysis.
rLCN2 protein
Recombinant mouse LCN2 protein has been prepared as previously described (33). In brief, rLCN2 protein was expressed as a GST fusion protein in the BL21 strain of E. coli, which does not synthesize siderophore. The protein was purified using glutathione-Sepharose 4B beads (Amersham Biosciences), followed by elution with thrombin or glutathione. For iron and enterochelin loading, a 5-fold molar excess of ion-saturated enterochelin (EMC Microcollections) was mixed with purified rLCN2 protein.
Statistical analysis
All data were presented as means ± SD from three or more independent experiments. Statistical comparison between different treatments was done by either the Student t test or one-way ANOVA with Dunnett’s multiple comparison tests using the GraphPad Prism program (GraphPad Software). Differences with a p value <0.05 were considered statistically significant.
Results
Establishment and characterization of apoptosis-resistant variant of microglial cells
We have previously shown that inflammatory activation of cultured microglia induces a self-regulatory apoptosis. To investigate the molecular mechanism(s) of the self-regulatory apoptosis of activated microglia, we selected and characterized apoptosis-resistant variants of BV-2 mouse microglial cells. Because NO is the major cytotoxic mediator in the apoptosis of activated microglia, variant clones of BV-2 microglial cells that were resistant to the NO toxicity were selected. After long-term culture of microglial cells in the presence of NO donor SNP, a variant clone of microglial cells that were resistant to the NO cytotoxicity was established and named BV-LS13. The NO-induced cell death in the parental BV-2 cells was due to apoptosis as determined by nuclear morphology, TUNEL staining, and induction of mitochondrial permeability transition (data not shown). The BV-LS13 was completely resistant to the NO-induced apoptosis, as opposed to parental cells (Fig. 1⇓). We next compared the global gene expression profile of the parental cells and the variant cells by cDNA microarray analysis (Fig. 2⇓C). One of the genes that showed a significant change in the mRNA expression levels in the BV-LS13 cells was lcn2: >5-fold decrease in the lcn2 mRNA expression has been detected in the BV-LS13 cells by the microarray analysis. The decrease in the lcn2 expression in the BV-LS13 cells was confirmed by RT-PCR and Western blot analysis (Fig. 2⇓, A and B). The expression of ATFx has been also investigated, because it has been previously reported that lcn2 and ATFx are oppositely regulated in the process of hemopoietic cellular apoptosis (30, 49). A significant change in ATFx expression between the parental and variant cells was not detected. The expression of serpinb6b, STAT1, and IRF-1 was also assessed for comparison, because the differences in the expression of these genes were shown in the microarray analysis. In the current study, we have focused on lcn2.
Establishment of the variant of BV-2 microglial cells that is resistant to NO-induced apoptosis. An apoptosis-resistant variant of microglial cells (BV-LS13) was selected after treatment of parental cells (BV-2) with 0.3 mM SNP for 26 days. The variant (BV-LS13) showed a complete resistance to NO toxicity. The cell viability was assessed by MTT assay. The results in this and all similar experiments were repeated several times and one representative done in triplicates is shown. The results are mean ± SD (n = 3).
Down-regulation of lcn2 expression in the apoptosis-resistant variant of microglial cells. Comparison of global gene expression profile between the parent cells and the variant using microarray analysis revealed a marked down-regulation of lcn2 expression in the variant BV-LS13 cells. This was confirmed by RT-PCR (A) and Western blot analysis (B). Differential expression of other genes (ATFx, serpinb6b, STAT1, and IRF-1) was also evaluated for comparison. Treatment of BV-LS13 cells with SNP for 8 h did not affect lcn2 expression. The expression of serpinb6b (A), STAT1, and IRF-1 (B) was modestly increased in BV-LS13 cells compared with BV-2 cells, while ATFx expression did not show a significant difference between BV-2 and BV-LS13 cells. Differential expression was based on the scatter plot of microarray analysis (C). Log-transformed spot intensity was plotted (M vs A scatter plot), normalized, and further analyzed as previously described (45 ). y- and x-axis indicate log2R/G and 1/2(log2RxG), respectively, where R and G are normalized signal intensity of Cy5 and Cy3, respectively. The spots corresponding to the genes investigated are indicated by the arrows.
Proapoptotic activity of lcn2 in microglia
As the expression of lcn2 was down-regulated in the apoptosis-resistant microglia and lcn2 has been previously involved in the cell death and survival, we hypothesized that the apoptosis-resistant phenotype of the BV-LS13 cells may be due to the marked down-regulation of proapoptotic lcn2 expression. To test this hypothesis, we used three different approaches: 1) stable overexpression of lcn2; 2) knockdown of lcn2 expression using shRNA; and 3) treatment with rLCN2 protein. First, BV-2 cells overexpressing lcn2 were obtained by stable transfection. An increase in the lcn2 mRNA as well as protein expression in the stable transfectants (S1 and S3) was confirmed by RT-PCR and Western blot analysis (Fig. 3⇓A). Stable overexpression of lcn2 enhanced the sensitivity of microglial cells to NO donors such as SNP and SNAP (Fig. 3⇓B). A NO donor SNP releases iron along with NO (50). A small amount of iron may affect microglial viability. To exclude this possibility, the effect of SNP was confirmed by another NO donor SNAP, which does not release iron. Both SNP and SNAP gave rise to a similar result. The apoptosis-enhancing effects of lcn2 in the stable transfectants were assessed by MTT assay (Fig. 3⇓B) as well as annexin V staining followed by flow cytometric analysis (data not shown). Transfection of lcn2 also increased the sensitivity of microglia to other cytotoxic agents such as etoposide, H2O2, and staurosporine (data not shown), indicating that lcn2 indeed has proapoptotic activity. The lcn2 transfectants S1 and S3, however, showed a slight difference in their sensitivity to cytotoxic drugs, which cannot be explained currently. Second, stable knockdown of lcn2 expression by the transfection of BV-2 cells with the lcn2-specific shRNA construct decreased the NO sensitivity (Fig. 4⇓B). Knockdown of lcn2 expression in the stable transfectants was confirmed by Western blot analysis (Fig. 4⇓A). To make sure that the effect of lcn2 knockdown observed is not due to an off-target effect, a shRNA rescue experiment was performed (Fig. 4⇓C). rLCN2 protein (Fig. 5⇓C) rescued the lcn2 knockdown effects (Fig. 4⇓C). Additionally, the recombinant mouse LCN2 protein sensitized BV-2 microglial cells to NO donors, whereas LCN2 protein alone did not affect microglial cell viability (Fig. 5⇓A). A similar result was obtained in primary microglia cultures (Fig. 5⇓B). Because it has been previously shown that the proapoptotic effect of LCN2 is associated with iron metabolism (29), the effect of iron chelator DFO on the microglial cell viability has been investigated. It has been reported that LCN2 containing the iron complex of the bacterial siderophore donates iron to cells via the lcn2 receptor (lcn2/24p3R) (29). Internalization of LCN2 and its receptor leads to the uptake of iron from the siderophore-iron complex. Donation of iron to the cell leads to a decrease in TfR1 expression and an increase in ferritin levels. In addition, donation of iron to the cell prevents apoptosis by decreasing the expression of the proapoptotic protein Bim. In contrast, LCN2 without an iron complex binds to lcn2/24p3R and is internalized into the cell, and then a putative intracellular mammalian siderophore iron complex becomes bound to LCN2, which is subsequently released from the cell by exocytosis. Depletion of iron from the cell results in the up-regulation of the proapoptotic molecule, Bim, which leads to apoptosis. DFO alone was modestly toxic to microglia cells (20 μM, 98.2 ± 1.2% viability; 100 μM, 86.3 ± 1.3% viability), and did not significantly affect the NO donor (SNP or SNAP)-induced microglial cell death (data not shown). Moreover, the LCN2-induced apoptosis sensitization was abolished by the concurrent addition of siderophore-iron complex (Fig. 5⇓D). Concentrations of LCN2, DFO, and siderophore-iron complex used in the present study are comparable to those in the previous report (29). These results indicate that the proapoptotic activity of LCN2 in microglia may be related with the iron transport and metabolism: apo-LCN2 may be proapoptotic by depleting intracellular iron, while holo-LCN2 containing the iron complex of siderophore may not be.
Apoptosis-enhancing effect of lcn2 in microglia. BV-2 microglial cells with increased lcn2 expression were selected by stable transfection of lcn2 cDNA. The increased lcn2 expression in the stable transfectants of lcn2 (S1 and S3) compared with the cells transfected with an empty vector (E1) was confirmed by RT-PCR and Western blot analysis (A). BV-2 microglial cells with increased lcn2 expression showed an enhanced sensitivity to the NO donor (SNP or SNAP)-induced apoptosis as determined by MTT assay (B) and annexin V-binding assay (data not shown). The lcn2 transfectants (S1 and S3) or the empty vector transfectant (E1) were exposed to the indicated concentrations of SNP or SNAP for 24 h, and then cell viability or apoptosis was evaluated by MTT assay (B) or flow cytometric analysis (data not shown). Values represent mean ± SD. ∗, Statistically significant differences from the empty vector transfectant (E1) (B) (p < 0.05). Results are the representative of three independent experiments.
Decreased apoptotic sensitivity in the lcn2-knockdown cells. Stable knockdown of lcn2 expression by transfection of BV-2 cells with lcn2-specific shRNA construct decreased the apoptotic sensitivity compared with the scrambled shRNA transfectant. Knockdown of lcn2 expression was confirmed by Western blot analysis of the transfectants (A). The viability of transfectants was measured by MTT assay after treatment with the indicated concentrations of SNP or SNAP for 24 h (B). The lcn2 knockdown effect was rescued by rLCN2 protein treatment (C). The rLCN2 protein reversed the decrease in the apoptotic sensitivity. Values are mean ± SD (n = 3). ∗, Statistically significant differences from the scrambled shRNA transfectant (p < 0.05).
rLCN2 protein sensitized microglia to apoptosis. Addition of rLCN2 protein increased the sensitivity of microglia to the NO toxicity. The apoptosis-enhancing effect of LCN2 protein was observed in both BV-2 cells (A and D) and primary microglia cultures (B). GST-fused LCN2 protein was expressed in BL21 cells (C, left), which was then cleaved by thrombin to release the pure LCN2 protein (C, right). The GST protein was run on the same gel for comparison. In contrast to apo-LCN2 protein, LCN2 protein complexed with iron and siderophore did not exert the apoptosis-enhancing effect (D). Values are mean ± SD (n = 3). Asterisks indicate statistically significant differences from the control treated with the same NO donor concentrations in the absence of LCN2 (∗, p < 0.05; ∗∗, p < 0.01) (A, B, and D).
Expression and regulation of lcn2, lcn2 receptor (lcn2/24p3R), and Bcl-2 family proteins in microglia
Previously, lcn2 has been proposed as an acute phase protein (40), and the expression of lcn2 has been shown to be modulated by inflammatory stimuli in macrophages (40, 51, 52). Thus, we sought to determine how the expression of lcn2 is regulated by inflammatory or toxic stimuli in microglia. The expression of lcn2 was strongly enhanced by LPS, serum withdrawal, PMA, IFN-γ, and calcium ionophore A23187 (Fig. 6⇓A). The secretion of LCN2 was also increased by LPS, serum withdrawal, and PMA as determined by Western blot analysis of conditioned medium of microglia cultures (Fig. 6⇓B). The results indicate that microglial expression and secretion of lcn2 may be increased under inflammatory condition in the CNS, as is in macrophages in the periphery. Recently, the presence of lcn2 receptor (lcn2/24p3R) that mediates the lcn2-induced apoptosis has been reported (29). The expression of this lcn2 receptor was detected in microglia (Fig. 6⇓C). Bcl-2 family proteins have been previously implicated in the effects of lcn2 on the cell death and survival (28, 29, 31). In particular, up-regulation of proapoptotic Bim expression was essential for the apoptosis-inducing effects of lcn2 (29). However, none of the proapoptotic Bcl-2 family proteins tested was significantly influenced by the lcn2 overexpression (E1 vs S1 or S3) with or without SNP treatment (Fig. 6⇓D). Most importantly, the expression of Bim was not changed by lcn2 transfection at mRNA (BimEL mRNA expression assessed by RT-PCR; data not shown) or protein levels (Fig. 6⇓D). These results suggest that the regulation of proapoptotic Bcl-2 family proteins may not be directly involved in the apoptosis-sensitizing effects of lcn2. Nevertheless, it is yet possible that the proapoptotic effect of lcn2 on microglia may be mediated through the newly identified lcn2 receptor (lcn2/24p3R).
Expression of lcn2, lcn2 receptor (lcn2/24p3R), and Bcl-2 family proteins and their regulation by inflammatory signals and other stimulants in microglia. The expression of lcn2 was increased by LPS (100 ng/ml), serum withdrawal for 8 h (SW), PMA (100 μg/ml), IFN-γ (50 U/ml), or calcium ionophore A23187 (5 μM). Treatment with SNP (0.5 mM), gangliosides mixture (G-mixture; 50 μg/ml), thrombin (10 μg/ml), ATP (3 mM), or forskolin (10 μM), however, did not significantly affect the lcn2 expression levels. BV-2 microglial cells were treated with the indicated stimuli for 24 h (A, upper), 8 h (A, lower), or were incubated under serum-free condition for 8 h (SW), and then LCN2 protein was detected by Western blot analysis (A). Secreted LCN2 protein was detected by Western blot analysis of culture media after the similar treatment for 8 h (B). RT-PCR analysis revealed that BV-2 and BV-LS13 cells as well as primary microglia cultures express lcn2 receptor (24p3R), which has been recently shown to mediate the apoptotic effect of lcn2 (C). The lcn2 receptor (24p3R) expression was not detected in the reaction without reverse transcriptase (data not shown). The expression of proapoptotic Bcl-2 family proteins such as BIM, BAX, and BAD was not significantly influenced by the lcn2 up-regulation or SNP treatment (D). lcn2 receptor expression was detected by RT-PCR analysis (C), and Bcl-2 family proteins were detected by Western blot analysis after treatment of BV-2 cells, empty vector (E1), or lcn2 (S1 and S3) transfectants with 0.5 mM SNP for 8 h (D). Either Ponceau S staining or α-tubulin detection was done to confirm the equal loading of the samples.
Deramification of microglia by lcn2
Stable expression of lcn2 in BV-2 microglia cells not only increased their sensitivity to apoptotic signals, but also induced a morphological change that reflected amoeboid transformation of microglia upon activation (Fig. 7⇓). A similar morphological change was also observed in primary microglia cultures treated with rLCN2 protein (Figs. 8⇓ and 9⇓). LPS has been previously shown to induce microglial apoptosis through TLR4 (53) and to cause microglial deramification (4, 47, 54). Activation of microglia with LPS treatment similarly induced deramification under the current condition (Fig. 8⇓).
Induction of morphological change by the forced expression of lcn2. Stable expression of lcn2 (S1 and S3) induced the deramification of BV-2 microglial cells. Parental BV-2 cells and the empty vector-transfectant (E1) are shown for comparison (original magnification, ×200). Scale bar, 25 μm.
The effect of purified rLCN2 protein on microglial morphology. Addition of rLCN2 protein (10 μg/ml) induced the morphological change in primary microglia cultures in a manner similar to the lcn2-expressing BV-2 cells (B). Treatment with LPS (100 ng/ml) also induced microglial deramification (C). Primary microglia cultures were either left untreated (A) or treated for 24 h with LCN2 protein or LPS, and then cellular morphology was examined under phase contrast microscope. The percentage of deramification was then quantitated (D). Several randomly chosen fields were examined to obtain the percentage of ramified cells, which were defined as the cells harboring the processes that are longer than the diameter of the cell body. Results are one representative of more than three independent experiments. GST (10 μg/ml) was used as a control protein (photo not shown). Original magnification, ×100 (A–C). Scale bar, 50 μm.
Comparison of LCN2 protein with other stimuli with respect to the effects on microglial morphology. After treatment of primary microglia cultures for 24 h with culture media, rLCN2 protein (10 μg/ml), ATP (3 mM), forskolin (10 μM), or A23187 (5 μM), cellular morphology was examined by isolectin B4 staining. Afterward, the percentage of deramification was quantitated as in Fig. 8⇑ (A). Representative cellular morphologies after the isolectin B4 staining of untreated or LCN2-treated microglia are shown (original magnification, ×100) (B). Scale bar, 50 μm.
Relationship between microglial deramification and apoptosis-prone phenotype
We next examined whether the LCN2-induced microglial deramification is related with the apoptosis-sensitive phenotype. We have used other signaling mediators or reagents that are previously known to induce microglial deramification in vitro, such as ATP, forskolin, and calcium ionophore A23187. As expected, all these stimuli induced the deramification of primary microglia cultures (Fig. 9⇑). These stimuli also increased the NO-induced cell death of microglia to varying degree (Fig. 10⇓). However, only calcium ionophore A23187 enhanced lcn2 expression (Fig. 6⇑A), suggesting that an increase in the intracellular calcium may lead to the microglial deramification possibly through lcn2 up-regulation. The effect of calcium ionophore A23187 was also assessed after the knockdown of lcn2 expression to further examine the role of lcn2. The lcn2 knockdown reversed the apoptosis-enhancing effect (Fig. 10⇓B, upper) as well as the deramification-inducing effect of A23187 (A23187 treatment in the scrambled shRNA transfectant, 65.07% ramification; A23187 treatment in the lcn2-specific shRNA-transfectant, 99.76% ramification, when the percent ramification of untreated control cells was set to 100%). Knockdown of lcn2 expression before and after the A23187 treatment in the lcn2 shRNA transfectant was confirmed by Western blot analysis (Fig. 10⇓B, lower). ATP and forskolin seem to exert their effects independently of LCN2. Moreover, ATP, forskolin, and calcium ionophore showed a modest cytotoxic effect on microglia at the concentration that induced the morphological change (Fig. 10⇓A), suggesting that the mechanisms of action of these stimuli may be different from that of LCN2. Cytotoxicity of ATP, forskolin, and calcium ionophore was not detected in the earlier studies, because the microglial morphology was examined on top of confluent astrocytes (47) or within 30 min to 2 h after treatment (55); at this time point, the cell death might not yet have occurred. In the current study, microglial viability and morphology was mainly assessed after 24 h (Figs. 8–10⇑⇑⇓). The morphological change also took place at 2 h (data not shown). Treatment of microglia with LCN2 alone induced the morphological transformation without apparent cytotoxicity under this condition. Taken together, these results suggest that amoeboid transformation of microglia and their apoptotic sensitivity are closely associated with each other and that LCN2 carries out a dual function via a unique mechanism in the regulation of microglial apoptosis and morphology.
Apoptosis-enhancing effects of LCN2, ATP, forskolin, and A23187 in microglia. BV-2 cells (A), lcn2 shRNA-transfected BV-2 cells (B), or primary microglia cultures (C) were exposed to LCN2 proteins (10 μg/ml) or other stimuli (ATP, 3 mM; forskolin, 10 μM; A23187, 5 μM) in the presence of 0.5 mM SNP for 24 h, and then the cell viability or apoptosis was determined by MTT assay (A and B, upper) or FITC-conjugated annexin V staining followed by counting of FITC-labeled cells under fluorescent microscope (C). Knockdown of lcn2 expression in the lcn2 shRNA-transfected BV-2 cells was confirmed by Western blot analysis (B, lower). The expression of lcn2 was assessed after treatment with calcium ionophore A23187 (5 μM) for 8 h. Bars represent mean ± SD values determined from three independent experiments (A and B) or the percentage of apoptotic FITC-labeled cells, which represents three independent experiments (C). ∗, Statistically significant differences from the treatment with SNP alone (A) or significant differences between the two treatments (B).
Discussion
In this study, we report that microglial secretory protein LCN2 plays a dual role in determining morphological and functional fate of activated microglia (Fig. 11⇓). LCN2 appears to sensitize activated microglia to apoptosis and it also induces deramification of microglia. This was supported by the following results: 1) the expression of lcn2 was down-regulated in the apoptosis-resistant microglia; 2) transfection of lcn2 cDNA conferred the apoptotic sensitivity; 3) shRNA-mediated knockdown of lcn2 expression diminished the apoptotic sensitivity, which was rescued by rLCN2 protein; 4) rLCN2 protein mimicked the proapoptotic effect of lcn2 cDNA transfection; and 5) lcn2 cDNA transfection and rLCN2 protein induced deramification of microglia. Based on these results, it is speculated that activated microglia may secrete LCN2 proteins, which act in an autocrine manner to induce morphological transformation of microglia (Fig. 11⇓). At the same time, secreted LCN2 proteins may sensitize activated microglia to apoptotic signals, so that activated microglia can be easily eliminated by apoptosis as a self-regulatory mechanism. LCN2 protein is constitutively expressed in microglia and its expression is strongly enhanced by inflammatory stimulation. The secreted LCN2 protein may feed back to microglia, and may induce the apoptosis-prone phenotypic change in the microglia which have already been stimulated by inflammatory signals. It is also possible that secreted LCN2 protein may influence neighboring cells (astrocytes, neurons, and endothelial cells) in the brain. In the inflammatory sites in the CNS, the number of microglia may increase by proliferation or inward migration. Thus, the secretion of LCN2 may be a part of an autoregulatory mechanism by which activated microglia in the inflammatory sites could be safely removed (Fig. 11⇓). This is in agreement with a previous report, where the role of secreted LCN2/24p3 in the termination of the immune response has been proposed (30). Previously, it has been reported that the production of RBC is also under the control of LCN2 in a similar manner (56). LCN2 acted as an autocrine factor that negatively regulates RBC production by inducing apoptosis and inhibiting differentiation.
Schematic diagram showing the dual role of lcn2 in the apoptosis and deramification of microglia. Once microglia are activated by inflammatory stimuli, LCN2 protein is secreted. Released LCN2 protein feeds back to microglia to induce morphological as well as functional changes. LCN2 induces deramification of microglia, and it also renders microglia more sensitive to apoptotic signals, which may provide a basis for the self-regulatory elimination of activated microglia in vivo. Based on the studies using other deramification-inducing stimuli, the morphological change of microglia appears to be closely related with the phenotypic change to apoptosis-prone microglia.
In the healthy brain, microglia have a typical ramified morphology consisting of a small cell body and long processes with secondary branching. Changes in the microglial morphology are observed in a wide range of the CNS pathologies such as brain injury, ischemia, infection, autoimmunity, and neurodegenerative diseases (7, 8, 9). Under these pathological conditions, ramified microglia transform into amoeboid microglia, with their processes retracted and their cell body enlarged. Such a morphological transformation is associated with microglial activation and is induced in vitro by a variety of stimuli including LPS, IFN-γ, laminin, ATP, etc. (4, 47, 54, 55, 57, 58). Nevertheless, the mechanism underlying the morphological transformation from ramified microglia into amoeboid microglia remains unknown. In the present study, we have unexpectedly found that LCN2, one of the secreted proteins of activated microglia, plays a central role in the morphological changes. Upon transfection of BV-2 microglial cells with lcn2 cDNA, such a morphological transformation occurred. This was confirmed by the treatment of primary microglia cultures with rLCN2 protein. Our results indicate for the first time that a single secretory protein of microglia actually mediates their own deramification. A similar type of morphological change has been also observed after the transient expression of lcn2 in NIH3T3 fibroblast cells (J. Lee and K. Suk, unpublished results), indicating that the effect of lcn2 on the cellular morphology may not be limited to microglia.
Having shown that lcn2 plays a dual role in microglial apoptosis and deramification, our next question was how these two effects of lcn2 are related with each other. To answer this question, we have used other stimuli that are known to induce the morphological change of microglia (47, 55, 57). Calcium ionophore, ATP, and forskolin all induced the deramification of microglia, and concurrently enhanced the sensitivity of microglia to the NO-induced apoptosis. These results indicate that microglial deramification is closely associated with the apoptotic sensitivity. The morphological change may be sufficient to increase the apoptotic sensitivity. This speculation, however, needs to be further tested using other conditions where microglial deramification takes place. Calcium ionophore exerted a unique effect on microglia in that it induced the microglial deramification with a concurrent induction of LCN2 expression, which further supported that lcn2 is responsible for microglial deramification. ATP and forskolin, however, did not enhance LCN2 expression, and yet induced the deramification and apoptotic sensitivity, suggesting that 1) ATP and forskolin may induce microglial deramification independently of LCN2 and 2) the microglial morphology is the major factor in determining cell survival and death.
The results reported in this study were obtained using microglial cells cultured in vitro, which may not be the same as brain microglia in vivo. Functional significance of lcn2 in the microglial apoptosis and morphological change in vivo has yet to be determined. The disruption of the blood-brain barrier following CNS injury in vivo can lead to the recruitment of blood-borne monocytes that are difficult to distinguish from the brain-resident microglia (54, 59). Thus, the results on microglial morphology either in vitro or in vivo alone have to be interpreted with caution. Recently, genome-wide expression profiling has shown that LCN2 is up-regulated in brain tissue after focal cerebral ischemia (60, 61). Whether microglia are the cellular source of LCN2 in brain, and whether the microglia-derived LCN2 controls their own apoptosis and morphology in vivo, needs further investigation.
Proapoptotic activity of LCN2 is controversial. LCN2-induced apoptosis seems to occur in a cell type-specific fashion. A previous study even reported the role of LCN2 as a survival factor (32). Differences between mouse and human LCN2 in terms of proapoptotic activity have been also reported (62). Devireddy et al. (30) has shown that most leukocytes are sensitive to LCN2-mediated apoptosis, while nonleukocytic cell lines are not. Among many primary cells of leukocytic origin tested, only macrophages as an exception were resistant to the LCN2-mediated apoptosis (30). As microglia are the CNS equivalent of macrophages, they were also resistant to the LCN2-mediated apoptosis as demonstrated in the current study. LCN2 alone did not induce apoptosis of microglia. LCN2 merely increased the sensitivity of microglia to other apoptogenic stimuli. Thus, as least in macrophages and microglia, it can be concluded that LCN2 is proapoptotic. It should be, however, noted that molecular mechanisms underlying the apoptosis-inducing effect and the apoptosis-sensitizing effect of LCN2 may be different. Under the condition where LCN2 induced apoptosis, it increased the expression of proapoptotic member of Bcl-2 family proteins such as Bim through the recently identified cell surface receptor for LCN2 (lcn2/24p3R). On the contrary, such a change in the expression of Bcl-2 family proteins was not observed in microglia, where LCN2 merely increased the apoptotic sensitivity. Also, phosphorylation of BAD at serine 136 was not detected (data not shown). Although the LCN2 receptor (lcn2/24p3R) is also expressed on microglia, it is possible that unidentified LCN2 receptors may be present on microglia whose proapoptotic pathways are not directly associated with the modification of Bcl-2 family protein expression. It should be also noted that lcn2-induced apoptosis occurred in the cells that overexpressed lcn2/24p3R (29). The level of lcn2/24p3R may influence the fate of microglia; it may determine whether LCN2 actually causes apoptosis or merely facilitates the apoptotic processes.
LCN2 has been implicated in mesenchymal-epithelial transition. LCN2 is an endogenous epithelial inducer (33) and stimulates the epithelial phenotype of transformed cells suppressing invasion and metastasis (63, 64). LCN2 also promotes tubulogenesis by regulating epithelial morphogenesis (48). The expression of E-cadherin was, however, not detected in primary microglia cultures by RT-PCR even after 40 cycles of amplification (data not shown). Also, there was no change in its expression after LPS or LCN2 treatment, arguing against the role of LCN2 in the mesenchymal-epithelial transition of microglia. It is not currently understood how the role of LCN2 as an epithelial inducer is related with the proapoptotic and amoeboid-transforming effects of LCN2 found in the present study. Nevertheless, what these studies have in common is that LCN2 has an important role in determining cellular morphology in addition to its role in the cell survival and death.
In conclusion, we present evidence that a secretory protein LCN2 is involved in the morphological transformation and apoptosis of activated microglia (Fig. 11⇑). The amoeboid transformation of microglia may be closely related with their vulnerability to apoptosis. Thus, activated microglia in vivo may secrete LCN2 protein, which induces morphological transformation and facilitates the self-regulatory apoptotic elimination. However, the exact molecular mechanisms underlying the proapoptotic activity of LCN2 and the morphological changes of microglia need to be elucidated in the future studies.
Acknowledgments
We thank N. Merillon for technical expert assistance and V. Cerundolo for critically reading the manuscript.
Disclosures
The authors would like to mention that the Institut Curie has decided to set up a biotech company partly based on this technology. This work has not been sponsored by this company and no author of this manuscript will have any responsibility in this company. No material described in this study will be developed and commercialized by this company. However, Ludger Johannes and Eric Tartour have been invited to acquire stocks (5550 Euros) in this company as coinventors of this technology.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 This work was supported by the Neurobiology Research Program from the Korea Ministry of Science and Technology and by Grant R01-2006-000-10314-0 from the Basic Research Program of the Korea Science and Engineering Foundation. S.L., J.L., and S.K. were supported by the Brain Korea 21 Project in 2006. J.-Y.P. and K.S. were the recipients of the Korea Research Foundation Grant funded by the Korean Government (MOEHRD, Basic Research Promotion Fund) (KRF-2006-005-J04202; KRF-2006-311-E00045).
↵2 Address correspondence and reprint requests to Dr. Kyoungho Suk, Department of Pharmacology, School of Medicine, Kyungpook National University, 101 Dong-In, Joong-gu, Daegu, 700-422, Korea. E-mail address: ksuk{at}knu.ac.kr
↵3 Abbreviations used in this paper: LCN2, lipocalin 2; SNAP, S-nitroso-N-acetylpenicillamine; SNP, sodium nitroprusside; DFO, deferoxamine mesylate; shRNA, short hairpin RNA.
- Received March 6, 2007.
- Accepted June 22, 2007.
- Copyright © 2007 by The American Association of Immunologists