Phagocytosis of bacteria is enhanced in macrophages undergoing nutrient deprivation Wim Martinet 1, *, Dorien M. Schrijvers 1, *, Jean-Pierre Timmermans 2 , Arnold G. Herman 1 and Guido R. Y. De Meyer 1 1 Division of Pharmacology, University of Antwerp, Belgium 2 Laboratory of Cell Biology and Histology, University of Antwerp, Belgium Nutrient deprivation triggers a variety of signaling events that enable energy conservation by cells. Among the different nutrient-sensing pathways, it is worth not- ing: (a) AMP-activated protein kinase, a metabolic stress sensor that is stimulated by elevated AMP ⁄ ATP ratios; (b) Per-Arnt-Sim kinase, which acts as a molec- ular sensor of oxygen, redox status, ATP and other indicators of the cellular metabolism; and (c) the hexosamine biosynthetic pathway that produces uri- dine 5¢-diphospho-N-acetylglucosamine as a substrate for O-N-acetylglucosamine transferase [1]. One energy conservation strategy that has attracted much attention is enhanced autophagocytosis, also known as macro- autophagy or simply autophagy [2]. This process is a catabolic pathway involving the engulfment and degra- dation of a cell’s own components through the lyso- somal machinery [3,4]. Autophagic signaling in response to nutrients is mainly relayed through the serine-threonine kinase mammalian target of rapa- mycin (mTOR). Indeed, mTOR is activated by nutri- ent-rich conditions, especially high levels of amino acids and insulin. Blocking mTOR function using rapamycin or its analogs mimics nutrient deprivation and triggers autophagy [1,4]. Keywords autophagy; heterophagy; p38 MAP kinase; scavenger receptor A; starvation Correspondence W. Martinet, Division of Pharmacology, University of Antwerp, Universiteitsplein 1, B-2610 Antwerp, Wilrijk, Belgium Fax: +32 3 820 25 67 Tel: +32 3 820 26 79 E-mail: wim.martinet@ua.ac.be *These authors contributed equally to this work (Received 28 November 2008, revised 28 January 2009, accepted 5 February 2009) doi:10.1111/j.1742-4658.2009.06951.x Phagocytosis represents a mechanism used by macrophages to remove pathogens and cellular debris. Recent evidence suggests that phagocytosis is stimulated under specific conditions of stress, such as extracellular pres- sure and hypoxia. In the present study, we show that amino acid or glucose deprivation caused an increase in the phagocytosis of heat-inactivated Escherichia coli and Staphylococcus aureus by macrophages, but not the uptake of platelets, apoptotic cells or beads. Increased phagocytosis of bac- teria could be blocked by phagocytosis inhibitors and was found to be dependent on p38 mitogen-activated protein kinase activity and scavenger receptor A. Although nutrient deprivation is a strong stimulus of auto- phagy, autophagosome formation was not critical for the uptake of bacte- ria because phagocytic clearance was not inhibited after down-regulation of the autophagy essential gene Atg7. Moreover, enhanced uptake of bacteria should not be considered as a general stress response because phagocytosis of bacteria was not stimulated after exposure of macrophages to the genotoxic agent camptothecin, heat (40 °C) or thapsigargin-induced endo- plasmic reticulum stress. Overall, the results obtained in the present study indicate that nutrient deprivation can stimulate macrophages to fight bacterial infections. Abbreviations AC, apoptotic cell; CHOP, C ⁄ EBP homologous protein; EBSS, Earle’s balanced salt solution; eGFP, enhanced green fluorescent protein; ER, endoplasmic reticulum; LPS, lipopolysaccharide; MAP, mitogen-activated protein; MARCO, macrophage receptor with collagenous structure; mTOR, mammalian target of rapamycin; NDRG1, N-myc downstream-regulated gene 1; PI, propidium iodide; PI3-kinase, phosphoinositide 3-kinase; PLT, platelet; RORa, RAR-related orphan receptor alpha gene; siRNA, small interfering RNA; SR-A, scavenger receptor A; TLR, Toll-like receptor. FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS 2227 Beside autophagocytosis, cells have the ability to internalize and degrade extracellular particles via het- erophagocytosis (hereafter referred to as phagocytosis) [5,6]. Lower organisms use phagocytosis for the acqui- sition of nutrients, whereas phagocytosis in metazoa primarily occurs in specialized cells, such as macro- phages and neutrophils, where it has evolved into an extraordinarily complex process. Phagocytosis by mac- rophages is critical for the uptake and removal of infectious agents and senescent or dying cells and is stimulated under specific conditions of stress, such as extracellular pressure [7] and hypoxia [8]. Several lines of evidence indicate that many nutrients regulate phag- ocytic activity in macrophages. For example, a choles- terol-rich diet inhibits macrophage phagocytosis by down-regulating plasma membrane fluidity and inter- ference with receptor movement [9]. By contrast, ascor- bate, as well as some trace elements, such as zinc, lead and cadmium, stimulate the phagocytic capacity of macrophages [10–12]. Furthermore, a diminished avail- ability of iron may impair the ability of phagocytic cells to kill ingested bacteria and fungi by down-regu- lating myeloperoxidase activity [13]. Recent evidence was provided showing that auto- phagy is essential for the phagocytosis of apoptotic corpses during embryonic development because the autophagic process contributes to the generation of engulfment signals in apoptotic cells (ACs) by main- taining cellular ATP production [14]. Moreover, engag- ing the autophagic pathway via Toll-like receptor (TLR) signaling also enhances phagosome maturation and the destruction of engulfed material [15]. It should also be noted that members of the phosphoinositide 3-kinase (PI3-kinase) family participate in autophago- some formation [16,17] and in phagocytosis through the delivery of membranes into extending pseudopodia [18]. These findings clearly indicate that the autophagic pathway is tightly linked to phagocytosis. In the pres- ent study, we examined whether nutrient deprivation, one of the main triggers of autophagy, stimulates the phagocytosis capacity of macrophages. Results Nutrient deprivation leads to an increase in the phagocytosis of heat-inactivated bacteria by macrophages in vitro Mouse J774A.1 macrophages were incubated in amino acid-free Earle’s balanced salt solution (EBSS) or con- trol medium supplemented with 10% fetal bovine serum for 6 or 24 h, followed by 1 h of incubation with fluorescently-labeled platelets (PLT), U937 ACs, carboxylated beads (0.1 or 1 lm in diameter) or heat- inactivated Escherichia coli bacteria. Flow cytometric analysis demonstrated that the uptake of PLTs, U937 AC or beads was not changed in EBSS-starved macro- phages versus control cells (Fig. 1A). By contrast, the uptake of E. coli bacteria was clearly increased after 6 and 24 h (Fig. 1A). To examine the potential role of TLR signaling in the enhanced clearance of bacteria, phagocytosis of beads was measured in the presence of 10 lgÆmL )1 lipopolysaccharide (LPS), which is a TLR ligand and important surface constituent of E. coli. Despite activation of macrophages by LPS, as deter- mined by nitrite measurements in the culture medium (12 ± 2 lm nitrite versus < 0.1 lm after 24 h of incu- bation with or without LPS, respectively), phagocytosis of beads was not stimulated. Increased phagocytosis of E. coli in macrophages undergoing EBSS-induced nutrient deprivation was confirmed by confocal microscopy (Fig. 1B,C). Optical slides in the three perpendicular axes showed that most bacteria were surrounded by macrophage cytoplasm in all dimen- sions (Fig. 1B). Moreover, enhanced uptake of bacteria was blocked by the phagocytosis inhibitor cytochala- sin D (Fig. 1D). We may therefore assume that the flow cytometry data truly reflect phagocytosis and not merely adherence of the bacteria to the macrophage surface. Apart from amino acid deprivation, glucose deprivation significantly increased the internalization of heat-inactivated E. coli, whereas incubation in serum-free medium had no effect (Fig. 2B). Similar findings were obtained after phagocytosis of the Gram-positive bacterium Staphylococcus aureus (Fig. 2B). Titration experiments demonstrated that glu- cose in the medium must be below 1 mgÆdL )1 to trig- ger enhanced bacterial uptake (not shown). Nutrient deprivation was also found to cause an increase in the extent of phagocytosis by peritoneal macrophages in vitro, as reflected by an increase in mean fluorescence of macrophages after phagocytosis of propidium iodide (PI)-labeled E. coli in amino acid deprivation versus control conditions (mean fluo- rescence: 116 ± 14 versus 48 ± 5, respectively; P < 0.001; unpaired Student’s t-test, n = 5). Enhanced phagocytosis of bacteria by starved J774A.1 macrophages is scavenger receptor A (SR-A) dependent Both SR-A and macrophage receptor with collagenous structure (MARCO) mediate the binding of unopson- ized bacteria in vitro and in vivo, and are suggested to play a pivotal role in bacterial clearance [19]. Although MARCO is the dominant receptor for unopsonized Phagocytosis of bacteria in starved macrophages W. Martinet et al. 2228 FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS bacteria on human alveolar macrophages [20], western blot analysis of J774A.1 lysate before and after starva- tion showed that these macrophages did not express MARCO (Fig. 2A). SR-A was overexpressed after amino acid or glucose deprivation, but also after serum withdrawal. Co-treatment of macrophages with anti-SR-A serum inhibited uptake of E. coli and S. aureus (Fig. 2B). Moreover, down-regulation of SR-A gene expression with gene-specific small interfer- ing RNA (siRNA) gave similar results (Fig. 2C), indi- cating that SR-A is essential for enhanced bacterial clearance. TLR4, also known as the LPS receptor, and the accessory molecule CD14 are involved in the phagocytosis of Gram-negative bacteria [21,22], but a potential role in the phagocytosis of E. coli after induction of autophagy is unlikely because TLR4 is down-regulated in starved J774A.1 cells (Fig. 2A). Moreover, phagocytosis of heat-inactivated E. coli was unaltered in nutrient-deprived peritoneal macrophages from TLR4 knockout mice (Fig. 2D). Starvation-induced autophagy is not involved in enhanced phagocytosis of bacteria by J774A.1 macrophages Because nutrient deprivation is a powerful inducer of autophagy, we next aimed to assess whether autophagy is induced in starved J774A.1 cells and whether this process affects the phagocytosis of bacteria. Transmis- sion electron microcopy revealed that both amino acid Fig. 1. Phagocytosis of heat-inactivated E. coli is enhanced in EBSS-treated J774A.1 macrophages. (A) Cells were incubated in serum-containing RPMI 1640 medium (con- trol) or treated with EBSS for 6 or 24 h. Subsequently, PLTs, U937 ACs, beads (1 lm) or E. coli bacteria (labeled with PI or CellTracker Red) were added to the culture medium. After 1 h of phagocytosis, the mean fluorescence of macrophages was measured by flow cytometry. *P < 0.05, ***P < 0.001 versus control (one-way ANOVA, followed by Dunnett’s or Dunnett’s T3 post-hoc tests, n = 8–16). (B) Confocal microscopy of CellTracker Green-stained J774A.1 macrophages after phagocytosis of PI-labeled E. coli for 1 h. Macrophages were incubated in serum-containing RPMI medium (control) or EBSS for 6 h prior to phagocytosis of E. coli. Bacteria were surrounded by macrophage cytoplasm in the three perpendicular optical sections. (C) Quantification of E. coli bacteria that adhered or were engulfed by J774A.1 macrophages after incubation in serum- containing RPMI medium (control) or EBSS for 6 h followed by 1 h of phagocytosis. ***P < 0.001 versus adherent (two-way ANOVA, n = 25). (D) J774A.1 cells were incubated in serum-containing RPMI medium (control) or EBSS in the presence or absence of cytochalasin D (2 l M) for 6 h prior to phagocytosis of fluorescently-labeled E. coli for 1 h. ***P < 0.001 versus without cytochalasin D, §§§ P < 0.001 versus control (two-way ANOVA, n = 9–15). W. Martinet et al. Phagocytosis of bacteria in starved macrophages FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS 2229 Fig. 2. Phagocytosis of heat-inactivated E. coli and S. aureus is enhanced in J774A.1 cells after nutrient starvation and is SR-A-dependent. (A) J774A.1 cells were incubated in RPMI 1640 medium supplemented with 10% fetal bovine serum (control), serum-free RPMI 1640 med- ium (without serum), EBSS [amino acid deprivation; without serum and amino acids (AA)] or glucose-free DMEM [without serum and glucose (glu)] for 6 h, followed by western blot analysis of macrophage receptors that are potentially involved in phagocytosis of heat-inacti- vated bacteria, such as MARCO, SR-A, TLR4 and CD14. In vitro translated MARCO cDNA served as a positive control for MARCO expres- sion. b-actin was used as a loading control. (B) Cells were incubated in RPMI 1640 medium supplemented with 10% fetal bovine serum (control), serum-free RPMI 1640 medium (serum deprivation), EBSS (serum and amino acid deprivation) or glucose-free DMEM (serum and glucose deprivation) for 6 h prior to phagocytosis of fluorescently-labeled E. coli or S. aureus for 1 h. Incubations were performed in the presence or absence of SR-A antibodies or nonspecific immunoglobulins (negative control antibodies). ++ P < 0.01, +++ P < 0.001 versus control (one-way ANOVA, followed by Dunnett’s test, n = 5–10); *P < 0.05, ***P < 0.001 versus without antibody and negative control anti- body (one-way ANOVA, followed by Bonferroni test, n = 5–10). (C) J774A.1 cells were transfected with SR-A-specific siRNA or siControl nontargeting siRNA. Three days after transfection, siRNA-treated cells were incubated in EBSS (serum and amini acid deprivation) for 6 h prior to phagocytosis of fluorescently-labeled E. coli for 1 h. **P < 0.01 versus control (unpaired Student’s t-test, n = 4). (D) Peritoneal mac- rophages from wild-type or TLR4 knockout (KO) mice were incubated in RPMI 1640 medium supplemented with 10% fetal bovine serum (control) or EBSS (serum and amino acid deprivation) for 6 h prior to phagocytosis of fluorescently-labeled E. coli for 1 h. ***P < 0.001 ver- sus control (two-way ANOVA, n = 7–9). Phagocytosis of bacteria in starved macrophages W. Martinet et al. 2230 FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS and glucose deprivation stimulated the formation of cytoplasmic vacuoles containing partially degraded cellular debris (Fig. 3A,B), a hallmark of autophagy. To evaluate whether these vacuoles represent autopha- gic vesicles, we assayed the intracellular formation of autophagosomes in RAW264.7 macrophages stably transfected with GFP-LC3 (Fig. 3C). Under nutrient- rich conditions, GFP-LC3 was found diffusely in the cytoplasm, with few punctuate dots. Amino acid depri- vation in EBSS induced the formation of autophago- somes that appeared as GFP-LC3 positive dots in the cytoplasm (Fig. 3C). Macrophages that underwent glucose starvation did not show significant changes in GFP-LC3 (Fig. 3C) compared to controls, despite the presence of large vacuoles in the cytosol early (6 h) after treatment (Fig. 3A). Instead, glucose-starved cells developed morphological features of necrotic death at later time points (24 h), such as organelle swelling and disruption of the plasma membrane (not shown), most likely as a result of ATP depletion and the lack of a supply of energy. Apart from GFP-LC3, we analyzed endogenous LC3 via western blotting. Because LC3 is poorly expressed in J774A.1 macrophages, the protein could be detected only in a reproducible way after transfection of LC3-encoding plasmid DNA. LC3 from transfected control cells, grown in serum-containing medium, was present mainly as the membrane-bound, autophagy-specific form LC3-II (Fig. 3D). Under EBSS conditions (i.e. without serum and amino acids), LC3-II markedly increased in the presence of the lyso- somal enzyme inhibitor NH 4 Cl, whereas it decreased in the absence of this inhibitor (Fig. 3D). These find- ings indicate that EBSS stimulates autophagosome formation, but that LC3-II is rapidly degraded by lysosomal hydrolases after fusion of autophagosomes with lysosomes. Neither amino acid-deprived cells, nor glucose-deprived J774A.1 cells underwent apoptosis because they did not demonstrate cleavage of caspase- 3, chromatin condensation, nuclear fragmentation or DNA fragmentation (not shown). Untreated macro- Fig. 3. Autophagy is induced in macrop- hages after amino acid deprivation, but not after glucose and ⁄ or serum deprivation. (A, B) Cells were incubated in RPMI 1640 med- ium supplemented with 10% fetal bovine serum (control), serum-free RPMI 1640 medium (without serum), EBSS [amino acid deprivation; without serum and amino acids (AA)] or glucose-free DMEM [without serum and glucose (glu)] for 6 h followed by trans- mission electron microscopy (A) and manual counting of vacuolated cells in transmission electron microscopy sections (B). Scale bar = 2 lm. **P < 0.01 versus control (one-way ANOVA, followed by Dunnett’s test, n = 5). (C) RAW264.7 macrophages expressing GFP-LC3 underwent nutrient deprivation for 2 h. Amino acid deprivation induced the formation of autophagosomes (arrowheads) in the cytoplasm. Scale bar = 20 lm. (D) Western blot analysis of LC3 in J774A.1 cells 24 h after LC3 trans- fection. Cells underwent nutrient deprivation in the presence or absence of the lysosomal enzyme inhibitor NH 4 Cl (10 mM) for 2 h. The LC3-II bands from three independent experiments were quantified and are shown as a percent of control (serum-containing medium without NH 4 Cl). *P < 0.05, **P < 0.01 versus control (one-way ANOVA, followed by Dunnett’s test, n = 3). W. Martinet et al. Phagocytosis of bacteria in starved macrophages FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS 2231 phages or cells incubated in medium without serum showed a normal cell morphology and did not undergo autophagy, apoptosis or necrotic death (Fig. 3A–D). Treatment of J774A.1 cells with the PI3-kinase inhibitors LY294.002 or 3-methyladenine blocked star- vation-induced phagocytosis of E. coli (Fig. 4A). To further test whether autophagosome formation is essential for the enhanced phagocytosis of bacteria, gene-silencing experiments were performed with siR- NA specific for the essential autophagy gene Atg7. Down-regulation of Atg7 gene expression with Atg7- specific siRNA was demonstrated at the mRNA level (91 ± 1% silencing after 24 h) and protein level (Fig. 4B), but not with siControl nontargeting siRNA. Vacuolization could not be stimulated in Atg7 siRNA transfected cells after EBSS treatment (Fig. 4B) and thus appear to be autophagy deficient. Atg7 silencing did not affect phagocytosis of E. coli (Fig. 4C), indi- cating that autophagosome formation is not critical for this process. Enhanced phagocytosis of bacteria by starved J774A.1 macrophages is p38 mitogen-activated protein (MAP) kinase dependent Amino acid or glucose deprivation, but not serum star- vation, substantially increased the phosphorylation of p38 MAP kinase (Fig. 5A). Inhibition of p38 MAP kinase during amino acid deprivation with the p38-specific inhibitor SB202190 blocked the enhanced clearance of heat-inactivated E. coli and S. aureus (Fig. 5B). Because p38 MAP kinase is potently and preferentially activated by a variety of environmental stresses [23], it is tempting to speculate that enhanced uptake of bacteria is a general stress response. To test this possibility, J774A.1 cells were exposed to heat (40 °C) or thapsigargin-induced endoplasmic reticulum (ER) stress for 6 h. Despite up-regulation of heat shock protein 70 and the ER stress marker C ⁄ EBP homologous protein (CHOP), respectively (Fig. 6A), Fig. 4. Autophagy is not essential for enhanced phagocytosis of heat-inactivated E. coli by amino acid-deprived J774A.1 macro- phages. (A) J774A.1 cells were incubated in serum-containing RPMI medium (control) or EBSS in the presence or absence of the autophagy inhibitors LY294.002 (LY; 50 l M) and 3-methyladenine (3-MA; 10 m M) for 6 h prior to phagocytosis of fluorescently- labeled E. coli for 1 h. ***P < 0.001 versus control; §§ P < 0.01, §§§ P < 0.001 versus EBSS-treated cells (one-way ANOVA, followed by Dunnett’s test, n = 12). (B) J774A.1 cells were trans- fected with Atg7-specific siRNA or siControl nontargeting siRNA. Silencing of Atg7 expression was evaluated after 0–3 days by western blotting. To evaluate the inhibition of autophagy, the number of vacuolated cells was counted via transmission electron microscopy after 6 h of EBSS treatment. ***P < 0.001 versus control (unpaired Student’s t-test, n = 5). (C) siRNA transfected cells were incubated in EBSS for 6 h prior to phagocytosis of fluorescently-labeled E. coli for 1 h. Differences between siControl and Atg7 siRNA-treated cells were not statistically significant (two-way ANOVA, n = 5). Phagocytosis of bacteria in starved macrophages W. Martinet et al. 2232 FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS phagocytosis of E. coli was not affected (Fig. 6B). How- ever, p38 was not hyperphosphorylated under these conditions (Fig. 6C) and, thus, we sought other stress situations that activated p38. Treatment of cells with the DNA damaging agent camptothecin caused strong phosphorylation of p38 (Fig. 6D), although without stimulating the uptake of heat-inactivated E. coli (Fig. 6E). We therefore concluded that enhanced phagocytosis of bacteria is not a general stress response and that p38 activation is required, but not sufficient, to increase the rate of bacterial clearance. To examine potentially important downstream path- ways linked to EBSS-induced starvation and p38 MAP kinase activation, a full genome microarray rep- resenting over 41 000 mouse genes or transcripts was probed with cDNA isolated from J774A.1 cells that were treated with EBSS, EBSS supplemented with the p38 inhibitor SB202190 or control medium supple- mented with 10% fetal bovine serum. The microarray data are available via the National Center for Bio- technology Information Gene Expression Omnibus (http://www.ncbi.nlm.nih.gov/geo/; accession number GSE14293). Among the genes that were differentially expressed, only N-myc downstream-regulated gene 1 (NDRG1) and RAR-related orphan receptor alpha (RORa), were up-regulated upon EBSS-induced star- vation. Up-regulation of NDRG1 and RORa was inhibited in the presence of SB202190. Microarray data were confirmed by real-time RT-PCR (Fig. 7A) and western blotting (Fig. 7B), with the exception of RORa, which could not be detected at the protein level (Fig. 7B). To examine whether NDRG1 and RORa are involved in the phagocytosis of bacteria, both genes were silenced via siRNA. Down-regulation of NDRG1 and RORa gene expression with gene- specific siRNA was demonstrated at the mRNA level (90 ± 1% and 84 ± 1% silencing of NDRG1 and RORa, respectively, after 24 h) and protein level (detectable only for NDRG1; Fig. 7C), but not with siControl nontargeting siRNA. Silencing of either gene did not affect the phagocytosis of E. coli (Fig. 7C), indicating that they do not play a critical role in this process. Fig. 5. Activation of p38 MAP kinase is required for enhanced phagocytosis of bacteria by J774A.1 macrophages. (A) J774A.1 cells were incubated in RPMI 1640 medium supplemented with 10% fetal bovine serum (control), serum-free RPMI 1640 medium (without serum), EBSS [amino acid deprivation; without serum and amino acids (AA)] or glucose-free DMEM [without serum and glu- cose (glu)] for 6 h, followed by western blot analysis of p38 and phospho-p38 (Thr180 ⁄ Tyr182) in crude cell lysate. (B) J774A.1 cells were incubated in serum-containing RPMI medium (control), serum-free RPMI (serum deprivation), EBSS (serum and amino acid deprivation) or glucose-free DMEM (serum and glucose deprivation) for 6 h in the presence or absence of the p38 inhibitor SB202190 (10 l M) prior to phagocytosis of fluorescently-labeled E. coli or S. aureus for 1 h. ***P < 0.001 versus without SB202190 (unpaired Student’s t-test, n = 5). W. Martinet et al. Phagocytosis of bacteria in starved macrophages FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS 2233 Discussion In the present study, we have demonstrated that the exposure of mouse J774A.1 cells or primary mouse peritoneal macrophages to nutrient deprivation (i.e. the withdrawal of amino acids or glucose in combina- tion with serum withdrawal) leads to a striking increase in bacterial phagocytosis. Furthermore, we have determined that the starvation effect on phagocy- tosis is relatively selective for bacteria because phago- cytosis of other extracellular particles, including platelets, ACs or carboxylated beads, was not stimu- lated. Phagocytosis of beads was not stimulated in the presence of LPS. We therefore consider that TLRs are not involved in the enhanced clearance of bacteria after nutrient deprivation. Furthermore, treatment of starved macrophages with beads resembling the size of bacteria (0.1 lm diameter instead of the regular 1 lm) did not improve their uptake, suggesting that the selec- tive engulfment of bacteria is not a consequence of particle size. However, internalization of beads in mac- rophages might occur via a phagocytosis-independent mechanism (e.g. emperipolesis) because the uptake of beads cannot be blocked with the phagocytosis inhibi- tor cytochalasin D [24]. Therefore, the possibility that the efficiency of phagocytosis depends on particle size cannot be entirely ruled out. Of note, nutrient depriva- tion is a strong stimulus of autophagy. Despite recent evidence linking the phagocytosis and autophagy path- ways [14,15], down-regulation of the autophagy essen- tial gene Atg7 in J774A.1 macrophages did not reveal a plausible connection between the enhanced phagocy- tosis of bacteria and the induction of autophagy. Nonetheless, starvation-induced phagocytosis of E. coli was blocked with LY294.002 or 3-methyladenine, which are two compounds that are widely used to inhi- bit autophagy. However, they act as PI3-kinase inhibi- tors [25] and, thus, may inhibit the uptake of bacteria by preventing the formation of pseudopodia, which is the primary role of class I PI3-kinase activity in Fig. 6. Enhanced phagocytosis of bacteria by J774A.1 macrophages is not a general stress response. (A–C) J774A.1 cells were incubated in serum-containing medium at 40 °C (heat) or in medium supplemented with 10 n M thapsigargin (to evoke ER stress) for 0–6 h followed by western blot analysis of heat shock protein 70 (Hsp70), ER stress marker CHOP, p38 and phospho-p38 (Thr180 ⁄ Tyr182). In addition, phago- cytosis of fluorescently-labeled E. coli was analyzed by flow cytometry using amino acid deprivation (EBSS) as a positive control. **P < 0.01 versus control (one-way ANOVA, followed by Dunnett’s test, n = 4). (D, E) J774A.1 cells were incubated in serum-containing medium sup- plemented with the genotoxic agent camptothecin (CT; 10 l M) for 6 h followed by western blot analysis of p38 and phospho-p38 (Thr180 ⁄ Tyr182). Phagocytosis of fluorescently-labeled E. coli was analyzed by flow cytometry using amino acid deprivation (EBSS) as a posi- tive control. **P < 0.01 versus control (one-way ANOVA, followed by Dunnett’s test, n = 5). Phagocytosis of bacteria in starved macrophages W. Martinet et al. 2234 FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS phagocytosis [18]. Moreover, even though the autopha- gic machinery enhances the maturation of phagosomes [15], it is not known whether it contributes to the engulfment of bacteria. Our findings suggest that the role of the autophagy machinery is limited to the maturation of the phagosome and that it does not participate in the initial internalization of bacteria. Macrophages express a broad spectrum of receptors that participate in bacterial recognition and internali- zation [5,6,26]. These receptors either recognize serum components (opsonins) that opsonize the bacteria (e.g. integrins, Fc- or complement-receptors) or directly rec- ognize molecular determinants on the bacterial surface. Because the incubation of bacteria with starved macro- phages took place in the absence of serum, phago- cytosis was opsonin-independent. Molecules of the scavenger receptor family have been implicated in the bacterial phagocytosis of unopsonized bacteria by mac- rophages. For example, SR-A can bind Gram-positive [27] and Gram-negative bacteria [28], including the species used in the present study (i.e. S. aureus and E. thinsp;coli). Studies with SR-A knockout mice revealed that SR-A plays an important role in host defense against bacterial infection; it enhances sensitiv- ities for S. aureus [29] and Listeria infection [30], as well as for LPS-mediated septic shock [31]. The results obtained in the present study indicate that SR-A was up-regulated in J774A.1 macrophages after nutrient starvation and that treatment of starved cells with an SR-A-specific antibody blocked the enhanced phagocy- tosis of E. coli and S. aureus after nutrient deprivation. It should be noted, however, that serum withdrawal also triggered SR-A expression without significant stimulation of bacterial phagocytosis. This finding indi- cates that SR-A is required, but not sufficient, for the enhanced uptake of bacteria after nutrient deprivation. Fig. 7. Expression of NDRG1 and RORa in J774A.1 macrophages after nutrient deprivation. J774A.1 cells were incubated in RPMI 1640 medium supplemented with 10% fetal bovine serum (control), serum-free RPMI 1640 medium (without serum), EBSS [amino acid deprivation; without serum and amino acids (AA)] or glucose-free DMEM [without serum and glucose (glu)] for 6 h in the presence or absence of the p38 inhibitor SB202190 (10 l M) prior to real-time RT-PCR (A) or western blotting (B). *P < 0.05 ver- sus without SB202190 (unpaired Student’s t-test, n = 4). (C) J774A.1 cells were transfected with NDRG1- or RORa-specific siR- NA or siControl nontargeting siRNA. Silencing of NDRG1 expres- sion was evaluated after 0–3 days via western blotting. Three days after transfection, siRNA-treated cells were incubated in EBSS for 6 h prior to phagocytosis of fluorescently-labeled E. coli for 1 h. Dif- ferences between siControl and siRNA-treated cells were not sta- tistically significant (two-way ANOVA, n = 5). W. Martinet et al. Phagocytosis of bacteria in starved macrophages FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS 2235 Analogous to SR-A, the scavenger receptor MARCO can adhere to Gram-positive and Gram-negative bacte- ria. MARCO is considered to be the major binding receptor for unopsonized bacteria on human alveolar macrophages [20], but was not detectable in J774A.1 macrophages, suggesting that its involvement in enhanced bacterial clearance, as reported in the present study, is unlikely. Similarly, the changes in TLR4 and CD14 cannot fully explain the enhanced uptake of bacteria by starved J774A.1 cells because these mole- cules only bind Gram-negative bacteria [21,22]. More- over, these receptors were not up-regulated after nutrient deprivation. Possibly, other receptors on the surface of J774A.1 cells are involved in binding and ⁄ or phagocytosis of bacteria [19]. The macrophage mannose receptor and lectin-like oxidized low-density lipoprotein receptor 1, for example, support the adhe- sion of Gram-positive and Gram-negative bacteria, but these receptors are not professional phagocytic recep- tors, and only bind receptors that require a partner to trigger efficient phagocytosis [32,33]. Although macrophage receptors are key proteins involved in the phagocytosis of bacteria [19], several lines of evidence suggest that other proteins may be equally important. For example, Anand et al. [8] dem- onstrated that hypoxia triggers phagocytosis of bacte- ria by macrophages in a p38 MAP kinase-dependent manner. Analogous with hypoxia, nutrient deprivation (both amino acids and glucose, but not serum) sub- stantially increased the phosphorylation of p38 MAP kinase. Inhibition of p38 MAP kinase activation by the p38-specific inhibitor SB202190 attenuated bacte- rial phagocytosis induced by nutrient deprivation. p38 MAP kinase regulates cell growth, cell differentiation, cell activation and cell death, and responses to inflam- mation and stress stimuli at the transcriptional and translational levels [34]. Many downstream substrates of p38 MAP kinase have been described, both in the cytoplasm and in the nucleus, that mediate these effects. Doyle et al. [35] reported that numerous TLR ligands specifically enhance the phagocytosis of bacte- ria through myeloid differentiation factor 88, interleu- kin-1 receptor-associated kinase-4 and p38, and that activation of this pathway is essential for the up-regu- lation of several scavenger receptors, including MARCO and SR-A, as well as lectin-like oxidized low-density lipoprotein receptor 1 and interstitial cell adhesion molecule-1. Hypoxia-induced phagocytosis of bacteria by macrophages occurs under the control of hypoxia-inducible factor-1a, whose expression is reversed after p38 inhibition [8]. In the present study, a microarray screening of the whole mouse genome did not reveal up-regulation of macrophage receptors or hypoxia-inducible factor-1a in starved J774A.1 cells, but led to the identification of two hypoxia-regulated genes (i.e. NDRG1 and RORa) that were strongly up-regulated at the mRNA level after amino acid or glucose deprivation, but not after serum withdrawal. The expression of both genes was controlled by p38 because up-regulation of NDRG1 and RORa mRNA was inhibited in the presence of SB202190. The overex- pression of NDRG1 was confirmed via western blot- ting; however, expressed protein from RORa was undetectable in J774A.1 cells even after starvation. Given that the uptake of bacteria is enhanced after hypoxia in macrophages [8] and that NDRG1 is up-regulated after hypoxia [36], it is tempting to specu- late that NDRG1 overexpression is associated with the induction of bacterial clearance. However, gene-silenc- ing experiments showed that neither NDRG1 nor RORa are essential for the phagocytosis of bacteria. Most likely, other factors may be expressed or released during starvation, which are undetectable via micro- array technology, and serve to augment phagocytosis independent of NDRG1 and RORa. Indeed, macro- phages are specialized immune cells that, once acti- vated, may release large amounts of different cytokines and ⁄ or reactive oxygen species. Secretion of interleu- kin-10 is a well-known stimulus for phagocytosis in human monocytes [37,38]. We previously demonstrated that J774A.1 macrophages secrete large amounts of the pro-inflammatory cytokines interleukin-6 and tumor necrosis factor a upon amino acid deprivation [39]; however, Anand et al. [8] demonstrated that treat- ment of macrophages with interleukin-1, tumor necro- sis factor a or reactive oxygen species (H 2 O 2 or NO) does not affect phagocytosis significantly. In summary, exposure of macrophages to amino acid or glucose deprivation selectively stimulates the phagocytosis of heat-inactivated Gram-positive and Gram-negative bacteria via an SR-A- and p38-depen- dent mechanism. From a clinical perspective, these results are promising with respect to the development of nutritional strategies for the treatment of patients who are infected with multi-resistant bacteria. Experimental procedures Cell culture Mouse J774A.1 macrophages and U937 monocytic cells were grown in RPMI 1640 medium (Invitrogen, Carlsbad, CA, USA) supplemented with antibiotics and 10% fetal bovine serum. RAW264.7 macrophages stably expressing GFP-LC3 [15] (a gift from M. Sanjuan, St Jude Children’s Research Institute, Memphis, TN, USA) were grown in Phagocytosis of bacteria in starved macrophages W. Martinet et al. 2236 FEBS Journal 276 (2009) 2227–2240 ª 2009 The Authors Journal compilation ª 2009 FEBS [...]... Franklin Lakes, NJ, USA) The acquisition threshold was set to include cells but to exclude debris Phagocytosis of bacteria in starved macrophages and remaining unbound particles from further analysis Data were analyzed using cell quest pro Software (BD Biosciences) Plasmid construction, transfection and in vitro translation The enhanced green fluorescent protein (eGFP) coding sequence was excised from... Zeiss, Jena, Germany), using an Argon laser (488 nm line) and a HeNe laser (543 nm line) with a bandpass emission filter (BP500-530) and a longpass emission filter (LP560), respectively Regions of interest containing individual macrophages were selected and Z-stacks were further analyzed using the ortho and gallery displays of the LSM510 imaging software After confocal imaging, adherent and engulfed bacteria. .. evaluation of a model for phagocytosis of cells undergoing apoptosis J Immunol Methods 287, 101– 108 25 Klionsky DJ, Abeliovich H, Agostinis P, Agrawal DK, Aliev G, Askew DS, Baba M, Baehrecke EH, Bahr BA, Ballabio A et al (2008) Guidelines for the use and interpretation of assays for monitoring autophagy in higher eukaryotes Autophagy 4, 151–175 26 Underhill DM & Ozinsky A (2002) Phagocytosis of microbes:... Sulahian TH, Ning YY, Yang Z, Pikkarainen T, Sankala M et al (2005) MARCO is the major binding receptor for unopsonized particles and bacteria on human alveolar macrophages J Immunol 175, 6058– 6064 21 Neal MD, Leaphart C, Levy R, Prince J, Billiar TR, Watkins S, Li J, Cetin S, Ford H, Schreiber A et al (2006) Enterocyte TLR4 mediates phagocytosis and translocation of bacteria across the intestinal barrier... 3-methyladenine (10 mm; Sigma), the phagocytosis inhibitor cytochalasin D (2 lm; Sigma) or the p38 MAP kinase inhibitor SB202190 (10 lm; Sigma) These inhibitors did not affect macrophage viability within the studied time frame In some phagocytosis experiments, the medium was supplemented with 10 lgÆmL)1 LPS from E coli O111:B4 (Sigma) to assess the potential role of TLR signaling Activation of macrophages. ..W Martinet et al serum-containing DMEM medium Peritoneal macrophages were isolated 5 days after injection of 2 mL of Brewer’s thioglycolate medium (Sigma, St Louis, MO, USA) into the peritoneal cavity of fasting C57Bl ⁄ 6 mice To generate ACs, U937 cells were incubated with 50 lm etoposide for 4 h Macrophages, U937 ACs and human blood platelet concentrates (kindly provided by the blood... and function in normal and disease states Carcinogenesis 29, 2–8 37 Capsoni F, Minonzio F, Ongari AM, Carbonelli V, Galli A & Zanussi C (1995) IL-10 up-regulates human monocyte phagocytosis in the presence of IL-4 and IFN-gamma J Leukoc Biol 58, 351–358 38 Lingnau M, Hoflich C, Volk HD, Sabat R & Docke WD (2007) Interleukin-10 enhances the CD14-dependent phagocytosis of bacteria and apoptotic cells... 370–379 11 Chvapil M (1976) Effect of zinc on cells and biomembranes Med Clin North Am 60, 799–812 12 Koller LD & Roan JG (1977) Effects of lead and cadmium on mouse peritoneal macrophages J Reticuloendothel Soc 21, 7–12 13 Chandra RK (1973) Reduced bactericidal capacity of polymorphs in iron deficiency Arch Dis Child 48, 864– 866 Phagocytosis of bacteria in starved macrophages 14 Qu X, Zou Z, Sun Q,... transfusion center of the University Hospital of Antwerp, Belgium) were fluorescently labeled with 10 lm CellTracker Green (Molecular Probes, Eugene, OR, USA) for 30 min at 37 °C For phagocytosis experiments, macrophages were washed in NaCl ⁄ Pi and incubated in RPMI 1640 medium with or without serum, DMEM without glucose or EBSS in the presence or absence of the autophagy ⁄ phagocytosis inhibitors LY294.002... product was BglII ⁄ HindIII digested and cloned in the similarly opened pN3 to produce pMuLC3 J774A.1 macrophages were transiently transfected using Fugene HD transfection reagent (Roche, Mannheim, Germany) at a ratio of 5 lL of reagent to 2 lg of plasmid DNA In vitro translation of mouse MARCO was performed using XbaI linearized pcDNA3MARCO plasmid (a gift from T Pikkarainen, Karolinska Institute, Stockholm, . respectively), phagocytosis of beads was not stimulated. Increased phagocytosis of E. coli in macrophages undergoing EBSS-induced nutrient deprivation was. 2D). Starvation-induced autophagy is not involved in enhanced phagocytosis of bacteria by J774A.1 macrophages Because nutrient deprivation is a powerful inducer of autophagy,