Triggering receptor expressed on myeloid cells 2 (TREM-2) is a cell surface receptor abundantly expressed on myeloid lineage cells such as macrophages and dendritic cells. It is reported that TREM-2 functions as an inflammatory inhibitor in macrophages and dendritic cells. However, the role of TREM-2 in bacterial killing remains unclear. This study explored the role of TREM-2 in bacterial eradication of Pseudomonas aeruginosa (PA), a Gram-negative bacterium which causes various opportunistic infections. Phagocytosis assay assessed by flow cytometry suggested that TREM-2 was not involved in the uptake of PA by macrophages, while bacterial plate count data showed that TREM-2 was required for macrophage-mediated intracellular killing of PA. Moreover, our results demonstrated that TREM-2 promoted macrophage killing by enhancing reactive oxygen species (ROS), but not nitric oxygen (NO) production. Treatment with N-acetylcysteine, a ROS scavenger, diminished the TREM-2-mediated intracellular killing of PA. To further investigate the underlined mechanisms of TREM-2-promoted bacterial killing, we examined the activation of downstream mitogen-activated protein kinases and PI3K/Akt pathway. Western blot data showed that silencing of TREM-2 inhibited phosphorylation of Akt, but not ERK, JNK or P38. In addition, pretreatment with PI3K active product PIP3 DiC16 reversed the elevation of intracellular bacterial load in TREM-2-silenced macrophages, while PI3K inhibitor wortmannin restored the decline of bacterial load in TREM-2-overexpressed macrophages. These data together suggested that the TREM-2-mediated bacterial killing is dependent on the activation of PI3K/Akt signalling, which may provide a better understanding of the host antibacterial immune defence.
Innate immunity is the first line of host defence against a variety of micro-organisms including bacteria, viruses and fungi . After infection, the invading microbes are recognized by pattern recognition receptors (PRRs) in host cells to activate innate immune response and sequentially initiate the adaptive immunity. As an important type of phagocytes and antigen presenting cells, macrophages play a critical role in the host immune defence system against a variety of bacteria. Studies have demonstrated that the immune and biological activities of macrophages are largely dependent on the activation of PRRs such as Toll-like receptors (TLRs)  and Nod-like receptors (NLRs) [3, 4] and downstream signalling pathways, such as mitogen-activation protein-kinase (MAPK) pathways  and PI3K-Akt pathways . Studies have demonstrated that TLRs can upregulated the production of reactive oxygen species (ROS) [7, 8] and reactive nitrogen species (RNS)  and eventually promotes the rate of bacterial elimination. However, little is known regarding the role of other PRRs.
As a novel PRR family, triggering receptors expressed on myeloid cells (TREMs) have recently emerged as important immune regulators . There are two major members in the TREM family, TREM-1 and TREM-2, which are mainly expressed on cells derived from the myeloid lineage. Although their ligands remained puzzled, studies using agonistic antibodies have unveiled the participation of TREM-1 and TREM-2 in modulating inflammatory responses. In regard to bacterial elimination, N'Diaye et al.  reported that overexpression of TREM-2 and its adaptor molecule DAP12 restores the impaired binding and uptake of a broad range of some Gram-positive and Gram-negative bacteria in DAP12 knockout bone marrow-derived macrophages (BMDMs). Charles et al.  also reported that in macrophages, high TREM-2 expression is associated with more ROS production, in response to Salmonella, but not zymosan. These reports indicate that TREM-2 may participate in the process of bacterial clearance, in a pathogen-dependent manner. Nonetheless, the role of TREM-2 in Pseudomonas aeruginosa (PA) elimination by host immune cells is still unclear.
Pseudomonas aeruginosa (PA) is a Gram-negative bacterium commonly existed in the environment and causes diverse opportunistic infections, such as chronic lung infection in cystic fibrosis patients, keratitis in contact-lense users, bacteraemia associated with severe burn injury . Studies have demonstrated that TREM-2 is constitutively expressed on alveolar macrophages , peritoneal macrophages , osteoclasts  and microglia , indicating that macrophage is the major cell source of TREM-2. It is reported that macrophages play a critical role in the host defence against PA infection such as PA-induced cystic fibrosis [17, 18] and ulcerative keratitis [19, 20]. Our previous study using a murine model of PA keratitis has demonstrated that TREM-2 is significantly enhanced in mouse corneas (mainly expressed on macrophages) after PA infection and promotes host resistance by reducing corneal inflammation and bacterial load .
A recent study demonstrates that phosphatidylinositol 3-kinase (PI3K) is recruited to TREM-2/DAP12 signal complex upon the interaction between PI3K regulatory subunit p85 and adaptor molecular DAP12, suggesting the involvement of PI3K in downstream TREM-2 signalling . As a family of lipid kinases, PI3K usually activates Akt, resulting in phosphorylation of a variety of downstream targets including IKK , mTOR , p21  and caspase 9 . Studies have demonstrated that PI3K/Akt signalling can modulate many cellular events, such as cytokine production, cell proliferation and apoptosis [27, 28]. Moreover, it is reported the PI3K/Akt pathway is required for killing of internalized pathogens, such as Neisseria gonorrhoeae . However, whether PI3K/Akt is involved in TREM-2-mediated bacterial clearance is still unknown.
In this study, we report that TREM-2 promotes macrophage-mediated killing of PA, by enhancing ROS generation and activating PI3K/Akt signalling pathway, which may shed lights on the interaction between the host and bacteria.
Materials and methods
PA strain 19660 was purchased from the American Type Culture Collection (ATCC, Manassas, VA, USA). Thioglycollate medium and Pseudomonas isolation agar were purchased from BD Difco Laboratories (Sparks, MD, USA). siRNA for mTREM-2 or appropriate scrambled control were obtained from Santa Cruz Biotechnology (Santa Cruz, CA, USA). PE-conjugated anti-TREM-2 Ab was purchased from R&D systems, and the isotype control PE-conjugated anti-rat IgG2b was from eBioscience (San Diego, CA, USA). Primary antibodies (Abs) against TREM-2, phosphorylated Akt (p-Akt), Akt and β-actin for Western blot were obtained from R&D systems (Minneapolis, MN, USA), Epitomics (Burlingame, CA, USA), Cell Signaling Technology (Danvers, MA, USA) and Sigma (St. Louis, MO, USA), respectively. Secondary Ab against sheep IgG was obtained from R&D systems, and secondary Abs against mouse and rabbit IgG were purchased from Bio-Rad (Hercules, CA, USA). Filmtracer Green Biofilm (FTGB) was from Invitrogen (Carlsbad, CA, USA). ROS scavenger N-acetylcysteine (NAC) was from Sigma. PI3K inhibitor wortmannin was purchased from Invitrogen. PIP3 DiC16 was purchased from Echelon (Salt Lake City, UT, USA).
RAW264.7 cell culture and transient transfection
Murine macrophage-like RAW264.7 cells (purchased from ATCC) were cultured in DMEM medium supplemented with 10% (vol/vol) FBS, 1% penicillin–streptomycin and 1% l-glutamine (all from Invitrogen) at the permissive temperature of 37 °C, as described by others [30, 31]. According to the manufacturer's instruction, RAW264.7 cells were transiently transfected with siRNA for TREM-2 versus appropriate scrambled control siRNA, or TREM-2 expression vector (pORF9-mTREM-2) versus control vector (pORF9-mcs) (all from Invivogen) using Lipofectamine™ 2000 (Invitrogen).
Total RNA was isolated from cell pellets using TRIzol (Invitrogen) according to the manufacturer's recommendations and quantitated by using NanoDrop 2000C Spectrophotometers (ThermoScientific, Waltham, MA, USA). One microgram of total RNA was reversely transcribed to produce cDNA by using RevertAid First Strand cDNA synthesis kit (Fermentas, Beijing, China) as previously described . Amplification of cDNA was conducted with Taq polymerase (Takara, Dalian, China) as suggested by the manufacturer. Primer sequences used to amplify mTREM-2 are as follows: 5′- GGA GGA CCC TCT AGA TGA CCA AGA-3′ (Forward); 5′- AGG CCA GGA GGA GAA GAA TGG A-3′ (Reverse). Primer sequences used to amplify β-actin are as follows: 5′- GAT TAC TGC TCT GGC TCC TAG C-3′ (Forward); 5′- GAC TCA TCG TAC TCC TGC TTG C-3′ (Reverse). Five microlitres of final PCR product was analysed by electrophoresis (1.2% agarose gel with ethidium bromide). Bands were visualized under UV transillumination.
Analysis of TREM-2 protein levels by flow cytometry
RAW264.7 cells were transfected with TREM-2 siRNA or expression vector versus their respectively controls for 24 h and then harvested, resuspended in binding buffer and incubated with PE-conjugated anti-TREM-2 (R&D systems) or isotype control PE-conjugated anti-rat IgG2b (eBioscience), according to the manufacturer's instructions. Cells were stained for 30 min on ice, washed twice with 1 ml FACS buffer and then resuspended in 1% formaldehyde. Flow cytometry was performed using LSRFortessa Cell Analyzer (BD Biosciences, San Jose, CA, USA), and data were analysed with Flowjo Software (Tree Star, Ashland, OR, USA).
RAW264.7 cells were washed three times with ice-cold PBS and then lysed in the lysis buffer. Then, protein concentration of the supernatant was determined by Quick Start Bradford protein assay (Bio-Rad). Twenty micrograms of each sample was loaded, separated on 10% SDS-PAGE and then transferred to a supported nitrocellulose membrane (Pall Life Sciences, Ann Arbor, MI, USA). After blockage, blots were incubated overnight with the respective primary Abs at 4 °C, followed by incubation with appropriate HRP-conjugated secondary Abs at room temperature for 1 h, as described by others . Finally, blots were visualized with Plus-ECL (PerkinElmer, Shelton, CA, USA) according to the manufacturer's protocol. The intensity of each band was measured using Adobe Photoshop 7.0 software (Adobe Systems Inc., San Jose, CA, USA), and relative integrated density values (IDV) of each band were calculated by normalizing to the β-actin control.
Phagocytosis assay by flow cytometry
Phagocytosis was assayed by flow cytometry as described by others . Briefly, PA was incubated with Filmtracer Green Biofilm (FTGB, 1:50; Invitrogen) at room temperature for 30 min, protected from light and then gently rinsed with sterilized water. Cells were challenged with FTGB-stained PA at MOI 25. After 1-h incubation, cells were treated with gentamicin (300 μg/ml, 30 min) and washed three times with cold PBS to remove extracellular bacteria. Then, cells were collected and analysed using a Beckman Coulter EPICS XL/MCL (Beckman Coulter Inc., Shanghai, China) instrument.
Intracellular bacterial killing assay
Intracellular bacterial load was assessed by plate count as described before [35, 36]. Cells were challenged with PA at MOI 10. After 1 h, cells were treated with gentamicin at 300 μg/ml for 30 min to kill the extracellular bacteria. Then, cells were washed with PBS for three times and lysed with 0.1% Triton-X. Intracellular bacterial CFU were determined by plate count and reported as CFU per cell ± SEM.
Immunofluorescence experiments were performed as described before . Cells were grown on collagen-precoated glass coverslips in 24-well plates. After transient transfection with TREM-2-siRNA or vector versus their respective controls for 24 h, cells were infected with FITC-labelled PA for 1 h. After wash with PBS, cells were fixed with 4% paraformaldehyde followed by membrane permeabilization using 0.2% Triton X-100. Then, cells were blocked with 5% BSA and incubated with DQ-Red-labelled CD63 and finally viewed by confocal microscopy (Zeiss Axiovert, LSM710, Oberkochen, Germany).
ROS measurement by flow cytometry
ROS measurement was performed as described by others . After PA challenge, cells were incubated with a ROS-sensitive probe 2′, 7′-dichlorofluorescin diacetate (H2DCFDA; Invitrogen) at a final concentration of 10 mm and then collected and analysed using a Beckman Coulter EPICS XL/MCL instrument (Beckman Coulter Inc.). ROS levels were determined by the fluorescence of DCF, the deacetylated and oxidized product of H2DCFDA [38, 39].
Supernatant of each cell sample was collected at 24 and 36 h after PA challenge. Nitric oxide (NO) levels were determined by measuring its stable end product, nitrite, using a Griess reagent (Sigma) as described before . Results were expressed as the mean micromoles of nitrite per sample ± SEM.
Student's t-test or ANOVA was used to determine the statistical significance of assays. Analysis was performed using Prism 5.0 software (GraphPad Software, Inc., La Jolla, CA, USA). Data were considered significant at P < 0.05.
The efficacy of silencing or overexpressing TREM-2 by transient transfection
To investigate the role of TREM-2 in murine macrophage-like RAW264.7 cells, both ‘loss-of-function’ and ‘gain-of-function’ studies were used. RAW264.7 cells were transiently transfected with either TREM-2-siRNA versus scrambled control siRNA, or pORF9-mTREM-2 plasmid versus control vector. mRNA levels of TREM-2 after silencing and overexpression were measured by conventional PCR (Fig. 1A, B) and real-time PCR (Fig. 1C, D), respectively. Furthermore, protein levels of TREM-2 after silencing (Fig. 1E) and overexpression (Fig. 1F) were measured by flow cytometry. The results showed that at both mRNA and protein levels, TREM-2 was constitutively expressed in RAW264.7 cells and downregulated after transfection of TREM-2 siRNA (Fig. 1A, C, E), but upregulated by transfection of pORF9-mTREM-2 plasmid (Fig. 1B, D, F), confirming the efficacy of TREM-2 silencing and overexpression.
TREM-2 enhanced intracellular bacterial killing but not phagocytosis in PA-challenged macrophages
To explore whether TREM-2 was involved in macrophage-mediated uptake and killing of PA, RAW264.7 cells were assessed using phagocytosis and killing assays based on flow cytometry and plate count, respectively. Neither silencing (Fig. 2A) nor overexpression of TREM-2 (Fig. 2B) had any effect on the uptake of PA. While bacterial load (as indicated by CFU per cell) was significantly enhanced by approximately 40% in TREM-2-silenced versus control-treated RAW264.7 cells (Fig. 2C, P < 0.05), consistently, after TREM-2 overexpression, bacterial load was reduced by approximately 60% (Fig. 2D, P < 0.05). These data together indicated that TREM-2 enhanced bacterial killing of PA by macrophages, but was not involved in the process of phagocytosis.
Moreover, phagolysosome fusion was detected by measuring the colocalization of phagosomes containing PA (FTGB-stained, as shown in green) with CD63-positive lysosomes (as shown in red). Confocal microscopy results showed that neither silencing (Fig. 3A, C) nor overexpressing (Fig. 3B, D) TREM-2 had any effect on phagolysosome fusion, suggesting that TREM-2 did not affect the process of phagolysosome fusion.
TREM-2 promoted macrophage-mediated bacterial killing by elevating ROS but not NO production
To further explore the microbicidal mechanisms involved, respiratory burst was measured in TREM-2-silenced or TREM-2-overexpressed RAW264.7 cells after PA challenge. Results showed that silencing of TREM-2 significantly reduced ROS production (Fig. 4A), while overexpression of TREM-2 significantly enhanced ROS production during PA infection (Fig. 4B), indicating that TREM-2 enhanced PA-induced ROS production. However, NO levels (as determined by its stable end product nitrite) were unchanged after silencing (Fig. 4C) or overexpression of TREM-2 (Fig. 4D). Similarly, overexpression of TREM-2 significantly enhanced the expression of NADPH oxidase NOX2 (Fig. 4E, P < 0.05), but had no effects on iNOS expression (Fig. 4F), further confirming that TREM-2 promoted ROS but not NO production.
Furthermore, treatment with N-acetylcysteine (NAC), a ROS scavenger, significantly diminished the TREM-2-mediated intracellular killing of PA (Fig. 4G). These data together indicated that TREM-2 promoted macrophage-mediated killing of PA by elevating ROS but not NO production.
In addition, to explore whether TREM-2 promoted ROS generation in a stimuli-dependent manner, phorbol-12-myristate-13-acetate (PMA) was used as a chemical stimuli to induce ROS generation. Our results showed that both ROS production and NOX expression were unchanged in TREM-2-silenced (or TREM-2-overexpressed) versus control-treated RAW264.7 cells in response to PMA challenge (Fig. 5A–D), indicating that the role of TREM-2 on ROS generation is stimuli dependent. No change in iNOS expression was detected in either TREM-2-silenced or TREM-2-overexpressed cells after PMA challenge (Fig. 5E, F), which is consistent with the observation during PA infection.
TREM-2 modulated bacterial clearance via a PI3K/Akt signalling pathway
To further explore the signalling pathway involved in TREM-2-mediated bacterial killing, Western blot was used to test the activation of MAPKs and Akt, which are important kinases in the regulation of innate immune responses. RAW264.7 cells were transiently transfected with TREM-2- or control siRNA for 24 h, followed by PA infection at MOI 5 for indicated time points. Protein levels of phosphorylated and total Akt as well as ERK, JNK and P38 were measured by Western blot. As shown in Fig. 6A, B, phosphorylated levels of Akt as well as ERK, JNK and P38 were increased at 15 min post-infection, suggesting that all of these pathways were activated by PA challenge. Interestingly, while phosphorylated Akt (p-Akt) levels were dramatically decreased after silencing TREM-2 (Fig. 6A), no change in phosphorylation of ERK, JNK or P38 was detected between TREM-2-silenced versus control groups (Fig. 6B). These results together suggested that PI3K/Akt, rather than ERK, p38 or JNK signals, were involved in the downstream signalling of TREM-2.
To further explore whether PI3K/Akt participates in the TREM-2-mediated bacterial clearance, RAW264.7 cells were pretreated with PIP3 DiC16 (the PI3K product which can activate Akt) versus control vehicle and then transfected with TREM-2 siRNA or scrambled control, followed by PA infection. Plate count data showed that silencing of TREM-2 enhanced the number of viable bacteria phagocytosed by the macrophages at 1 h post-infection, while pretreatment with PIP3 DiC16 reversed the elevation of intracellular bacterial load in TREM-2-silenced RAW264.7 cells (Fig. 6C). On the other hand, we pretreated RAW264.7 cells with wortmannin (a PI3K inhibitor), then transiently transfected the cells with was TREM-2 expression vector or control vector and finally challenged with PA for 1 h. Consistently, overexpression of TREM-2 decreased intracellular bacterial load in RAW264.7 cells at 1 h after PA challenge, but this reduction was reversed by pretreatment with wortmannin (Fig. 6D). These data together suggested that PI3K/Akt signalling pathway was required in the TREM-2-mediated bacterial killing of PA.
As an opportunistic pathogen, PA often causes serious infections in immunocompromised individuals [41, 42] and is associated with corneal infection in contact-lens users . Numerous studies reported that PRRs play critical roles in immune resistance against PA infection [43-45]. Our previous study using a murine model of PA keratitis has demonstrated that TREM-2 promoted host resistance against PA infection by reducing bacterial load , while the underlined mechanism remains unclear. In the present study, we demonstrated that TREM-2 directly promotes macrophage-mediated intracellular killing of PA by activating PI3K/Akt.
Bacterial eradication is a complicated process involving several steps. First, phagocytes need to bind to the bacteria and then uptake the adhered bacteria by virtue of different phagocytic receptors. Finally, the phagocytosed bacteria undergo killing and digestion via the intracellular bactericidal system . Daws et al. found that TREM-2 binds to a variety of bacteria such as Escherichia coli and Streptococcus pyogenes, but not PA or Staphylococcus xylosus , suggesting that TREM-2 does not directly influence the binding of PA to macrophages. N'Diaye et al.  reported that exogenous expression of TREM-2 and the adaptor molecule DAP12 induced Chinese hamster overy cells to uptake Escherichia coli, Francisella tularensis and Staphylococcus aureus, but not zymosan particles, whereas in our study, no change in phagocytosis of PA was detected in RAW264.7 cells after silencing or overexpressing TREM-2, indicating that TREM-2 was not involved in the interalization process of PA in macrophages. It is reported that leucocyte-adhesion-deficient phagocytes which do not express CR3 fail to ingest 50% of the PA strains isolated from patients with cystic fibrosis, while CD14 is involved in the ingestion of these PA strains which are unable to be phagocytosed by phagocytes in the absence of CR3, suggesting that different PA strains may be phagocytosed by different phagocytic receptors . Therefore, we suppose that the controversy between N'Diaye's report and our observation is largely due to the specificity of phagocytes and pathogens.
Once bacteria are ingested, they are killed within the activated macrophages through either oxygen-dependent or oxygen-independent bactericidal system. Here, we observed that silencing of TREM-2 enhanced the bacterial load, while overexpression of TREM-2 reduced the number of viable bacteria within the macrophages, suggesting that TREM-2 promotes the macrophage-mediated intracellular bacterial killing of PA. However, the colocalization of phagosome containing PA with lysosome was comparable in TREM-2-silenced (or TREM-2-overexpressed) versus control-treated RAW264.7 cells, indicating that TREM-2 did not affect the process of phagolysosome fusion. We further tested the involvement of the oxygen-dependent bactericidal system, which consist of two major components, ROS and RNS. ROS, including various oxygen intermediates such as O2−, OH−, H2O2, are generated immediately after phagocytosis via respiratory burst . While for RNS, activated phagocytes express iNOS, which sequentially induces NO production for bacterial killing . Studies showed that TREM-2 modulates ROS and iNOS/NO production in a pathogen-dependent manner [51, 52]. Our data provide evidence that TREM-2 promotes ROS but not NO production in response to PA infection, while upon PMA challenge, TREM-2 did not affect the ROS production and NOX expression, indicating that the TREM-2-mediated ROS production is stimuli dependent. More importantly, treatment with ROS scavenger NAC suppressed the TREM-2-mediated killing of PA in macrophages, implicating the participation of ROS in the TREM-2-mediated regulation of PA clearance.
We further examined the downstream signalling pathway involved in TREM-2-mediated bacterial killing. It is reported that in macrophage-differentiated osteoclasts, TREM-2 activated Akt and ERK [22, 53]. While in dendritic cells, TREM-2/DAP12 stimulation is through activation of ERK, but independent of NF-κB and P38 stress-activated protein kinase . Studies have demonstrated that Akt is required in the killing of Neisseria gonorrhoeae  and Erk is involved in host defence against Klebsiella pneumoniae , whereas JNK is reported to inhibit macrophage killing of Staphylococcus aureus  and p38 is shown to attenuate macrophage antibacterial activities against Streptococcus pneumoniae . Our data showed that silencing of TREM-2 inhibited Akt phosphorylation, but had no effect on activation of ERK, P38 and JNK. It is reported that inhibition of PI3K/Akt signalling abrogated the uptake of PA strain PAK by Madin–Darby canine kidney (MDCK) or HeLa cells . In our study, pretreatment with wortmannin also suppressed the macrophage-mediated phagocytosis of PA strain ATCC 19660 (data not shown), but significantly enhanced the number of viable bacteria in TREM-2-overexpressed cells after 1-h incubation. These data together demonstrated that TREM-2 reduced bacterial load by activating PI3K-Akt axis, which is consistent with our previous study showing that TREM-2 promoted resistance against PA corneal infection through a PI3K/Akt pathway . In addition, Chatterjee et al.  reported that PI3K/Akt activation could trigger ROS production in a model of pulmonary ischaemia. However, in the process of TREM-2-mediated macrophage killing of PA, whether there is a link between ROS generation and PI3K/Akt activation still needs further investigation.
Overall, our study explored the role of TREM-2 in macrophage killing of PA, a pathogen leading to various opportunistic infections. This may provide potential strategy for the control of this public health problem.
This work was supported by grants from National Natural Science Foundation of China (31200662, 31370868, 81261160323), Guangdong Innovative Research Team Program (2011Y035, 2009010058), The 111 Project (No. B13037), Specialized Research Fund for the Doctoral Program of Higher Education of China (20100171110047, 20120171120064), Guangdong Natural Science Foundation (10251008901000013, S2012040006680), Guangdong Province Universities and Colleges Pearl River Scholar Funded Scheme (No. 2009), National Science and Technology Key Projects for Major Infectious Diseases (2013ZX10003001).