P2X₇ Receptor Regulates Internalization of Antimicrobial Peptide LL-37 by Human Macrophages That Promotes Intracellular Pathogen Clearance

The main families of antimicrobial peptides in mammals are defensins and cathelicidins. As the only endogenous cathelicidin in humans, LL-37 exhibits potent antimicrobial activities against a broad spectrum of pathogens, including bacteria, viruses, and parasites, and it possesses additional functions important for inflammation and modulation of the immune system (1, 2).

It was reported that the biological functions of LL-37 are mediated by several cell surface receptors, such as formyl peptide receptor (FPR)2/ALX (3), P2X₇ receptor (P2X₇R) (4), and epithelial growth factor receptor (5). Previous reports demonstrated that P2X₇R mediated LL-37–induced IL-1β production by human monocytes (4), IL-8 production (6), and PGE₂ production (7) by human gingival fibroblasts. Our recent study also showed that LL-37 promotes LTB₄ and TXA₂ production via P2X₇R (8). P2X₇R is a ligand-gated ion channel, which is highly expressed by cells of the hematopoietic lineage, and it mediates cell death, killing of infectious organisms, and regulation of inflammatory responses (9). This receptor exists within the plasma membrane as a large multimolecular complex, consisting of β-actin, integrin β2, three heat shock proteins (10), and nonmuscle myosin (11). Moreover, the channel protein pannexin (Panx)-1 can be activated by prolonged stimulation of P2X₇R (12).

Internalization of LL-37 by different human cells is related to different biological functions of LL-37. For example, LL-37–induced IL-8 expression in epithelial cells seems to rely on the internalization and localization of LL-37 to the perinuclear region (13). Furthermore, LL-37 internalization by immature human dendritic cells was related to cell maturation (14). Moreover, it was reported that LL-37 directs human monocyte differentiation into the M1 phenotype, which may also be linked to LL-37 internalization (15). In addition, LL-37 internalization was demonstrated in endothelial cells (16, 17). However, it is not clear how LL-37 is internalized by human cells and how internalized LL-37 couples with intracellular second messengers to exert its functions.

Endocytosis or uptake is characterized by internalization of molecules from the cell surface into internal membrane compartments, and vesicular trafficking can be divided into two main pathways: the classical, clathrin-mediated, endocytotic pathway and the nonclassical, clathrin-independent, but lipid raft–dependent, pathway (18). Clathrin-mediated endocytosis (CME) is the uptake of material into the cell from the surface using clathrin-coated vesicles, and this pathway encompasses the internalization of nutrients, Ags, growth factors, and receptors (19). CME is the most well-characterized mechanism for the entry of molecules into cells through early and late endosomes to lysosomes. Generally, sorting and recycling endosomes can be considered early endosomes that are discriminated on the basis of function. In sorting endosomes, cargo is sorted for recycling back to the plasma membrane (or the Golgi) via recycling endosomes or to lysosomes via late endosomes. The lysosome is a major degradation site for internalized material and cellular membrane proteins (20).

However, emerging evidence shows that clathrin-independent endocytosis also exists. One form of clathrin-independent endocytosis relies on cholesterol-rich membrane domains, such as lipid rafts and caveolae. Caveolae/lipid raft–dependent endocytosis is involved in multiple biological processes, including entry of virus and bacteria into host cells and internalization of sphingolipids, endothelin, growth hormone, and IL-2Rs (21). The scaffolding protein caveolin-1 was suggested as a key component in the formation of caveolae, because the lack of caveolin-1 in null mice leads to the absence of caveolae (22, 23).

We recently observed LL-37 internalization by human macrophages, a process functionally linked to eicosanoid biosynthesis (8). In this study, we investigated the mechanism of LL-37 internalization by human macrophages and the functional contributions of this process to intracellular pathogen clearance.

PMA, 2-ME, HEPES, RPMI 1640 medium, cytochalasin B (CytoB), KN-62, oxidized ATP (oxATP), BzATP, filipin III, dynasore hydrate, nystatin, (-)-blebbistatin, wortmannin, geldanamycin, chloropromazine (CLQ), chloroquine (CLQ), FIPI hydrochloride hydrate, RIPA buffer, RPMI 1640 medium, and LPS (from Salmonella enterica) were from Sigma-Aldrich (St. Louis, MO); U0126 and SB203580 were purchased from Tocris Bioscience (Bristol, U.K.); fluorescein-conjugated Staphylococcus aureus, FBS, and M-CSF were purchased from Life Technologies (Paisley, U.K.); Abs against clathrin, caveolin-1, and P2X₇R were from Santa Cruz Biotechnology (Santa Cruz, CA); and synthetic LL-37 (NH₂-LLGDFFRKSKEKIGKEFKRIVQRIKDFFRNLVPRTES-COOH), FAM- or TAMRA-conjugated LL-37, TAMRA-conjugated sequence-scrambled LL-37 (sLL-37;GLKLRFEFSKIKGEFLKTPEVRFRDIKLKDNRISVQR), and Panx-1 antagonist peptide ¹⁰Panx (WRQAAFVDSY) were from Innovagen (Lund, Sweden).

We tested all inhibitors used in our study, and none of them showed any significant toxicity to cells under our experimental conditions (data not shown).

The THP-1 cell line was purchased from American Type Culture Collection (Manassas, VA) and cultured in RPMI 1640 supplemented with 10% heat-inactivated FBS, HEPES (25 mM), 2-ME (0.05 mM), penicillin (100 U/ml), and streptomycin (100 μg/ml). THP-1 cell suspension was maintained at 0.3–0.8 × 10⁶ cells/ml, and cell differentiation was induced by PMA (10 ng/ml) for 48 h.

Human mononuclear cells and polymorphonuclear neutrophils (PMNs) were isolated from freshly prepared buffy coats (Karolinska Hospital Blood Bank, Stockholm, Sweden) by gradient centrifugation on Ficoll-Paque PREMIUM (GE Healthcare Bio-Sciences, Uppsala, Sweden). Differentiation of human monocytes to macrophages was achieved by plating mononuclear cells in cell culture plates for 2 h; unbound cells were washed away with PBS, followed by cell culture over 7 d in RPMI 1640 supplemented with 10% FBS (heat inactivated), 25 mM HEPES, M-CSF (50 ng/ml), penicillin (100 U/ml), and streptomycin (100 μg/ml).

PMA-differentiated THP-1 (dTHP-1) cells or human monocyte-derived macrophages (HMDMs) were pretreated or not with different inhibitors, according to the experimental designs, followed by incubation with fluorescence-labeled LL-37 for a certain time period, as indicated. The cells were collected and washed three times with PBS. The mean fluorescence intensity (MFI) of cells was measured by a FACSCalibur and analyzed using BD CellQuest software (both from BD Biosciences, Franklin Lakes, NJ).

dTHP-1 cells or HMDMs were cultured in Lab-Tek II chamber slides (Thermo Fisher Scientific, Rochester, NY). The cells were incubated with fluorescence-labeled LL-37 for 1 h and washed three times with PBS. For analysis of colocalization of LL-37 and lysosome, Golgi, or lipid raft, LysoTracker, CellLight Golgi-GFP, or lipid raft marker cholera toxic subunit-B (CT-B) was added at 1 h, 16 h, or 30 min, respectively, before the cells were washed. For analysis of colocalization of LL-37 and clathrin, caveolin-1, or P2X₇R, the cells were fixed and permeabilized after incubation with fluorescence-labeled LL-37. Thereafter, the cells were incubated with Abs against clathrin, caveolin-1, or P2X₇R overnight at 4°C, followed by incubation with Cy3-conjugated goat anti-rabbit IgG or Alexa Fluor 647 donkey anti-goat IgG (Jackson ImmunoResearch, West Grove, PA) at room temperature for 1 h. Alexa Fluor 647–conjugated anti–EEA-1 Ab (MBL, Woburn, MA) was used to directly label early endosomes. After staining, the slides were mounted in VECTASHIELD Antifade Mounting Media with DAPI (Vector Laboratories, Burlingame, CA) and imaged with a confocal microscope (Olympus FV1000; Olympus, Tokyo, Japan or Zeiss LSM 700; Carl Zeiss, Jena, Germany).

After treatment with 10 μg/ml LL-37 for a certain time period, dTHP-1 cells were washed three times with PBS and lysed with RIPA buffer, together with a complete protease inhibitor mixture (Roche Diagnostics, Mannheim, Germany). Samples with equal amounts of protein were resolved by SDS-PAGE using NuPAGE Novex 4–12% Bis-Tris gels (Life Technologies) and electroblotted onto a nitrocellulose transfer membrane, according to the manufacturer’s instructions (Life Technologies). Membranes were soaked for 2 h in 0.05% T-TBS (20 mM Tris-HCl [pH 7.5], with 0.5 M NaCl and 0.05% Tween 20) containing 3% fat-free dry milk and incubated overnight at 4°C with primary Abs. Immunoreactivity was detected with a secondary anti-mouse or anti-rabbit Ab conjugated with HRP (GE Healthcare Bio-sciences). Immunoreactive bands were visualized with an ECL Western blot detection system (GE Healthcare Bio-sciences).

THP-1 cells were grown and maintained as described above. Lentiviral constructs with short hairpin RNA (shRNA) directed against human P2X₇R were obtained from Sigma-Aldrich. Lentivirus was prepared by transient cotransfection of HEK293T cells with lentiviral shRNA vectors (pLKO.1-puro), together with third-generation packaging constructs (pMDLg/pRRE+pRSV-Rev+pMD2.G). Control cells were obtained by transfection with lentivirus obtained from the pLKO.1-puro nontarget shRNA control vector (Sigma-Aldrich SHC002). Cells were infected for 16 h, followed by repeated washes to remove unincorporated virus particles. Stable knockdown (KD) cells were selected by culture with puromycin (10 μg/ml) for 2 wk. The efficiency of KD was verified by RT-PCR and Western blot analysis.

dTHP-1 cells (1 × 10⁶) were incubated with 10 μg/ml LL-37 at 37 or 4°C for a certain time period, according to the experimental design, and washed three times with PBS. Cells were incubated with equal numbers of S. aureus strain B5381 at 37°C with shaking. After 2 h, the cells were lysed using ice-cold water and vortexed for 30 s. Thereafter, the lysates were serially diluted and plated onto agar plates. The following day, viable bacteria were counted (CFU). The relative CFU was calculated as treatment group/control group.

The level of cellular reactive oxygen species (ROS) was measured with a DCFDA Cellular ROS detection assay kit, according to the manufacturer’s instructions (Abcam, Cambridge, U.K.). Briefly, dTHP-1 cells were cultured and differentiated in black, clear-bottom, 96-well plates (Corning, Corning, NY). After treatment, cells were washed with 1× buffer and stained with 100 μl diluted DCFDA for 45 min at 37°C in the dark. Subsequently, the solution was removed, and 100 μl 1× buffer was added. The fluorescence intensity of the cells was read using fluorescence spectroscopy (POLARstar OPTIMA, BMG Labtech, Ortenberg, Germany), with an excitation wavelength of 485 nm and an emission wavelength of 520 nm to represent ROS activity. Relative ROS activity was calculated as the value for treatment group/control group.

The level of lysosome formation was represented by fluorescence intensity of the cells after lysosomes were labeled with LysoTracker Probes (Life Technologies). dTHP-1 cells were cultured and differentiated in black, clear bottom, 96-well plates. One hour before the treatments ended, 100 nM LysoTracker Red DND-99 Probe was added to the plates. The medium was removed, and the fluorescence intensity of the cells was read and calculated with fluorescence spectroscopy (POLARstar OPTIMA) using an excitation wavelength of 544 nm and an emission wavelength of 610 nm.

Freshly isolated PMNs were pretreated with CytoB (10 μM) and CaCl₂ (2 mM) for 5 min at 37°C and then treated with LTB₄ (100 nM) for 1 min at 37°C. Subsequently, the supernatants were collected by spinning down the cells at 300 × g for 10 min, and dTHP-1 cells were incubated with the collected supernatants at 37°C for 1 h. Internalized LL-37 in dTHP-1 cells was detected by Western blot analysis.

Results are presented as mean ± SD. Differences between the means were evaluated using the Student t test or one-way or two-way ANOVA plus the Tukey multiple-comparison test. A p value < 0.05 was considered statistically significant.

When HMDMs were incubated with FAM-labeled LL-37, the fluorescence intensity of the cells increased in a manner that was dependent on LL-37 dose and incubation time (Fig. 1A, 1B). The same pattern was observed in dTHP-1 cells (Fig. 1C). However, the fluorescence intensity of dTHP-1 cells decreased significantly when they were treated with the endocytosis inhibitor CytoB before incubation with FAM-labeled LL-37 (Fig. 1D), which suggested that internalization of LL-37 by human macrophages is an endocytotic process.

Furthermore, LL-37 internalization was significantly inhibited at 4°C compared with incubation at 37°C, as assessed by FACS analysis (Fig. 1E) and confocal microscopy (Fig. 1F). Also, dTHP-1 cells did not take up sLL-37 (Fig. 1G, 1H), suggesting that the internalization is sequence specific. HMDMs displayed a similar pattern to dTHP-1 cells (data not shown).

We next investigated the endocytosis pathways involved in LL-37 uptake by human macrophages. The inhibitors for both CME (CLP, CLQ, and dynasore) and caveolae/lipid raft endocytosis (nystatin and filipin) were used. FACS analysis showed that all three inhibitors of CME significantly suppressed LL-37 internalization by HMDMs and dTHP-1 cells (Fig. 2A, 2B). In contrast, nystatin also dramatically suppressed LL-37 endocytosis by human macrophages, whereas filipin only showed inhibitory effects on HMDMs (Fig. 2A) but not dTHP-1 cells (data not shown). The results from Western blot analysis confirmed that the inhibitors of clathrin- and caveolae/lipid raft–dependent pathways blocked LL-37 internalization by dTHP-1 cells (Fig. 2C). Furthermore, when dTHP-1 cells were preincubated with nystatin and dynasore together, an additive inhibitory effect on the subsequent LL-37 internalization was detected (Fig. 2D).

In addition, we observed colocalization of internalized LL-37 with CT-B (lipid raft marker) (Fig. 2E, top panels) and caveolin-1 (Fig. 2E, middle panels) by confocal microscopy. Colocalization of internalized LL-37 and clathrin also was detected (Fig. 2E, bottom panels). Together, these results suggested the involvement of both clathrin-dependent and caveolae/lipid raft–dependent pathways in LL-37 endocytosis by human macrophages.

As detected by confocal microscopy, internalized LL-37 translocates further into the endosomal (Fig. 2F, top panels) and lysosomal (Fig. 2F, middle panels) compartments. In contrast, internalized LL-37 also seems to locate in the Golgi apparatus, as shown in Fig. 2F (bottom panels).

We found that both KN-62 (a specific P2X₇R inhibitor) and oxATP (an irreversible antagonist of P2X₇R) significantly suppressed LL-37 internalization by dTHP-1 cells, indicating that P2X₇R could be involved in the uptake process (Fig. 3A, 3B); similar results were detected in HMDMs (data not shown). In contrast, the specific antagonist of FPR2, WRW4, exhibited no effects on LL-37 internalization (Supplemental Fig. 1).

To further confirm the involvement of P2X₇R in LL-37 internalization, we generated P2X₇R-KD THP-1 cells by specific human P2X₇R-targeted shRNA. P2X₇R mRNA and protein expression was analyzed to verify P2X₇R deficiency in P2X₇R-KD THP-1 cells (Supplemental Fig. 2). Consistently, we observed a significantly lower level of internalized LL-37 in P2X₇R-KD cells compared with cells transfected with control vector (control) (Fig. 3C, 3D). In addition, the inhibitory effect of KN-62 on LL-37 internalization was almost completely abolished in P2X₇R-KD cells (Fig. 3E), which further confirmed the involvement of P2X₇R in LL-37 internalization.

Colocalization of internalized FAM–LL-37 and P2X₇R also was observed in dTHP-1 cells, but it was only partial (Fig. 3F). We also found that nystatin further suppressed LL-37 internalization by P2X₇R-KD dTHP-1 cells, whereas dynasore, the inhibitor of CME, did not (Fig. 3G). Furthermore, an LL-37/P2X₇R complex was primarily associated with clathrin when observed under a confocal microscope, whereas an association of LL-37/P2X₇R with caveolin-1 was rarely observed (Fig. 3I).

We next used specific pharmacological tools to delineate P2X₇R-associated molecules and signaling pathways in the LL-37 internalization process. BzATP (specific agonist of P2X₇R), ¹⁰PanX (peptide antagonist of Panx-1), and wortmannin (specific inhibitor for PI₃K) significantly inhibited LL-37 internalization into dTHP-1 cells, whereas (-)-blebbistatin (inhibitor of nonmuscle myosin II), FIPI (PLD isoform inhibitor), geldanamycin (inhibitor of hsp90), and U0126 and SB203580 (inhibitors of ERK and p38 MAPK) had no significant inhibitory effect (Fig. 3H). These results provided further evidence for the involvement of P2X₇R in LL-37 internalization by dTHP-1 cells and suggested that Panx-1 and the PI₃K signaling pathway may be associated with P2X₇R-mediated LL-37 internalization.

We demonstrated previously that the bactericidal activity of dTHP-1 cells is significantly increased after treatment with LL-37 for 24 h (24). In the current study, we found that the intracellular killing of S. aureus by dTHP-1 cells, treated with 10 μg/ml of LL-37 for only 2 h at 37°C, also was significantly enhanced (Fig. 4A). In contrast, 10 μg/ml of sLL-37 exhibited no effect (data not shown). However, when the same experiment was conducted at 4°C, a temperature at which almost no LL-37 was internalized by dTHP-1 cells (Fig. 1E, 1F), intracellular killing of S. aureus by dTHP-1 cells was not changed compared with control cells (Fig. 4A). This result indicated that LL-37 internalized by dTHP-1 cells could contribute to LL-37–induced intracellular bacterial killing.

To further address this issue, we also assessed LL-37–induced intracellular killing of S. aureus in P2X₇R-KD dTHP-1 cells, because P2X₇R-KD dTHP-1 cells exhibited less LL-37 internalization (Fig. 3C, 3D). Thus, when wild-type (WT; untransfected) dTHP-1 cells, control cells (sham-transfected), and P2X₇R-KD dTHP-1 cells were treated with 10 μg/ml of LL-37 for 2 h at 37°C, the subsequent intracellular survival of S. aureus in WT and control dTHP-1 cells was significantly suppressed, whereas this effect of LL-37 was abolished in P2X₇R-KD dTHP-1 cells (Fig. 4B).

In another experiment, we observed that when dTHP-1 cells were incubated with 10 μg/ml of LL-37 for 2 or 24 h, nearly the same amount of LL-37 was detected in cell lysates at these two time points (Fig. 4C), which suggested that LL-37 endocytosis became saturated after ∼2 h. However, if dTHP-1 cells were incubated with LL-37 for 2 h and then washed thoroughly with PBS, followed by incubation of the cells in LL-37–free medium for another 22 h, almost no internalized LL-37 was detected (Fig. 4C). This result indicated that internalized LL-37 could be degraded in dTHP-1 cells. In line with this result, the intracellular survival rate of S. aureus was much higher in dTHP-1 cells cultured in LL-37–conditioned medium for 2 h, followed by LL-37-free medium for 22 h, compared with dTHP-1 cells cultured in LL-37–conditioned medium for 2 h (Fig. 4D).

All of this evidence demonstrates that internalized LL-37 plays an important role in intracellular bacterial killing. In addition, results from the immunofluorescence study confirmed that internalized LL-37 colocalized with bacteria (S. aureus) in endosomes and lysosomes (Fig. 4E).

We next investigated potential mechanisms for how internalized LL-37 contributes to intracellular bacterial killing and found significantly increased ROS activity in LL-37–treated dTHP-1 cells compared with control cells (Fig. 5A). Moreover, the inhibitors of LL-37 internalization (dynasore, nystatin, and KN-62) significantly suppressed LL-37–induced ROS activity (Fig. 5B). Meanwhile, LL-37 stimulated much less intracellular ROS formation in P2X₇R-KD THP-1 cells than in control cells (Fig. 5C). Thus, internalization of LL-37 enhances ROS activity by dTHP-1 cells. We also confirmed that LL-37–induced bactericidal activity was diminished when ROS formation was inhibited (Fig. 5D), which indicated that LL-37–induced ROS production mediated LL-37–enhanced bactericidal activity.

As an intracellular antimicrobial effector, the lysosome plays a key role in pathogen clearance. In our experiments, internalized LL-37 colocalized with lysosomes (data not shown). Interestingly, LL-37 treatment also led to accumulated, highly concentrated intracellular lysosome formation (Fig. 5E). The enhanced lysosome formation in LL-37–treated dTHP-1 cells was confirmed by the significantly increased level of LysoTracker fluorescence intensity, as analyzed by fluorometry (Fig. 5F). Meanwhile, the inhibitors (dynasore and nystatin) that block LL-37 internalization also suppressed LL-37–induced lysosome formation (Fig. 5G). In addition, LL-37 triggered significantly less lysosome expression in P2X₇R-KD dTHP-1 cells than in control cells (Fig. 5H).

In the next experiment, we tested whether human macrophages could import PMN-derived LL-37. In a previous study, we showed that human PMNs are able to release a large amount of LL-37 in response to the proinflammatory lipid mediator LTB₄ (25). Thus, we collected the conditioned media from human neutrophils triggered with different treatments, and incubated dTHP-1 cells with these conditioned media. As shown in Fig. 6A, almost no LL-37 was detected in dTHP-1 cells cultured in normal culture medium. However, abundant LL-37 was noted in dTHP-1 cells cultured with the medium from LTB₄-treated PMNs that was rich in PMN-derived LL-37. Furthermore, when dTHP-1 cells were pretreated with different endocytosis inhibitors, their internalization of human PMN-derived LL-37 was suppressed (Fig. 6B). A similar experiment was performed with P2X₇R-KD dTHP-1 cells, which internalized significantly less PMN-derived LL-37 than did control cells (Fig. 6C). Together, these results demonstrate that dTHP-1 cells can engulf LL-37 exported from neighboring PMNs.

Antimicrobial peptides, including LL-37, are released by human PMNs, epithelial cells, and additional inflammatory cells during infection and certain inflammatory responses. The released peptides interact with the surrounding macrophages to enhance killing and clearance of bacteria (24), promote adhesion of classical monocytes (16, 17), and regulate the production of inflammatory mediators in different cell types (26–28); these responses typically are transduced by cell surface receptors, in particular FPR2/ALX (3). However, recent work by us (8) and other investigators (13–15) demonstrated that LL-37 can be engulfed by cells, leading to bioactions. In this study, we explored the mechanism(s) by which LL-37 is taken up by human macrophages and identified a novel role for internalized LL-37 in intracellular bacterial killing.

In this study, we present several lines of evidence demonstrating that P2X₇R plays an important role in LL-37 internalization by human macrophages. First, LL-37 endocytosis by human macrophages was suppressed by P2X₇R antagonists (KN-62 and oxATP). Furthermore, LL-37 internalization was significantly decreased in cells deficient in the receptor (P2X₇R-KD THP-1 cells). Colocalization of LL-37 and P2X₇R also could be detected on the cell membrane and inside the cells of LL-37–treated macrophages using confocal microscopy. Finally, pretreatment of human macrophages with BzATP, a specific P2X₇R agonist that elicits receptor internalization (29), dramatically suppressed the subsequent uptake of LL-37 (Fig. 3H).

P2X₇R is highly expressed in macrophages, microglia, and certain lymphocytes, and it mediates the influx of Ca²⁺ and Na⁺ ions, as well as the release of proinflammatory cytokines. P2X₇R is distinguished from other P2X family members by a very low affinity for extracellular ATP (mM EC₅₀) and its ability to trigger the induction of nonselective pores upon repeated or prolonged stimulation (30). It was reported that LL-37 enters human PBMCs independent of P2X₇R and FPR2 (31). However, this conclusion was based on experiments with PBMCs, whereas we used HMDMs and dTHP-1 cells. It was reported that differentiation of monocytes into macrophages greatly increases the expression and function of P2X₇R (32, 33). Gudipaty et al. (34) concluded that the function of P2X₇R can be modulated by receptor subunit density at the cell surface and ambient levels of extracellular Na⁺ and Cl⁻, which may prevent adventitious P2X₇R activation in monocytes until these proinflammatory leukocytes migrate to extravascular sites of tissue damage. Therefore, P2X₇R may play different and context-dependent roles in LL-37–related responses in monocytes and macrophages.

P2X₇R within the plasma membrane forms a large multimolecular complex containing associated molecules, such as β-actin, integrin β2, three heat shock proteins (10), and nonmuscle myosin (11). The dissociation between P2X₇R and the connected molecules results in diminished P2X₇R-induced effects. It also was reported that activation of P2X₇R can trigger PLD-, MAPK-, or PI₃K-mediated downstream signaling pathways (9). Therefore, we tested several inhibitors of the P2X₇R-associated molecules or signaling pathways and found that PI₃K signaling and Panx-1 seem to be involved in P2X₇R-mediated LL-37 internalization. It was reported that Panx-1 is a P2X₇R-associated protein and appears to be the large pore or is responsible for activation of the large pore of P2X₇R (12). Interestingly, Panx-1 also was required for the recognition and intracellular delivery of bacterial molecules and caspase-1 activation (35, 36).

It is unclear whether the specific domain(s) of P2X₇R or some coassociated adhesion molecule directly binds to or recognizes LL-37. Although P2X₇R is a well-known cation channel, it also acts as a scavenger receptor and is involved in the direct recognition and engulfment of apoptotic cells (37). We analyzed the binding of LL-37 to the cell membrane and found that it was recovered only in the Triton X-100 detergent fraction of cell membrane (data not shown), which suggests that LL-37 interacts with the plasma membrane via hydrophobic interactions. Moreover, it was demonstrated that LL-37 is oriented nearly parallel to the surface of zwitterionic lipid membranes (eukaryotic cell membranes) (38). Clearly, understanding the interactions between LL-37 and membranes requires further studies with other, more detailed biochemical, biophysical, and imaging techniques.

It was reported that endocytosis of LL-37/DNA complexes by mammalian cells involves noncaveolar lipid raft domains, as well as cell surface proteoglycans (39). However, the route by which LL-37 enters human cells has not been elucidated. In this study, we demonstrate that both the CME pathway and the caveolae/lipid raft pathway are involved in LL-37 internalization by human macrophages. The additive effect of the pharmacological inhibitors indicates that two pathways might separately mediate LL-37 internalization (Fig. 2D). A previous report showed that ATP-stimulated P2X₇R internalization occurs through the clathrin domain (29). Accordingly, we also identified that the LL-37/P2X₇R complex primarily existed associated with clathrin (Fig. 3I). In addition, the inhibitor of CME (dynasore) lost its effect on LL-37 internalization by P2X₇R-KD dTHP-1 cells, whereas an inhibitor of caveolae/lipid rafts (nystatin) still suppressed this response in P2X₇R-KD cells (Fig. 3G). Together, our results suggest that P2X₇R-mediated LL-37 internalization is primarily associated with CME, which is consistent with the fact that CME tends to be a receptor-mediated pathway (19). However, considering our recent report that LL-37 was internalized by human macrophages to induce COX-2 activation and downstream inflammatory lipid eicosanoid production via a P2X₇R-independent pathway (8) and our present observation that internalized FAM–LL-37 and P2X₇R were only partially colocalized, it is possible that pathways other than P2X₇R-mediated internalization are involved in LL-37 endocytosis.

Macrophages are frontline pathogen sensors and rapidly engulf pathogens and kill them. When encountering the pathogens at infectious sites, neutrophils immediately release their preformed granules containing antimicrobial agents, including LL-37. Thus, resident macrophages may be surrounded by high levels of LL-37 at infectious or inflammatory sites. Our data show that macrophages take up external cathelicidin to acquire higher bacterial killing activity and confirm that macrophages internalize LL-37 released from human neutrophils. We observed direct colocalization of internalized LL-37 and the phagocytized bacteria in endosomal and lysosomal compartments, which suggested that LL-37 might encounter and eliminate pathogens directly in these organelles. In line with our observations, a previous report demonstrated that macrophages also take up apoptotic neutrophils to acquire neutrophil granules for antimicrobial activity against intracellular pathogens (40).

We also found that LL-37–treated macrophages produce enhanced levels of intracellular ROS and accumulate lysosomes, which could facilitate intracellular bacterial killing. This ROS production and lysosome formation may well be linked to LL-37 internalization because endocytosis inhibitors suppressed LL-37–induced responses, and much less ROS production and lysosome formation was detected in P2X₇R-KD THP-1 cells than in control cells. We also found evidence suggesting that LL-37–induced ROS contributes to LL-37–promoted intracellular bacterial killing because inhibition of ROS synthesis abolished LL-37–promoted intracellular bacterial killing (Fig. 5D). However, it is not clear how internalized LL-37 links to ROS production. It was reported that LL-37 can induce ROS formation in neutrophils, most likely via NADPH oxidase activation (41), and CRAMP (the murine ortholog of LL-37) also induces ROS activities in neutrophils through an FPR-independent pathway (42). Interestingly, Alalwani et al. (43) found that, after stimulation with S. aureus, peritoneal neutrophils from CRAMP-deficient mice produce significantly lower levels of ROS than the cells from WT mice, which underscores the role of the endogenous peptide in this process. Moreover, it was reported that AMPs (including LL-37) synergize with anionic endotoxin and enhance ROS generation by macrophages. This synergistic enhancement of ROS production also was seen in vitro during conversion of xanthine to uric acid and H₂O₂ by xanthine oxidase, suggesting a direct effect of AMPs on the oxidase (44).

Actually, P2X₇R function has been linked to the ability of macrophages to handle pathogens (9). For example, P2X₇R activation stimulates macrophages to kill intracellular Mycobacterium tuberculosis (45, 46), Chlamydia spp. (47), Leishmania amazonensis (48), and Toxoplasma gondii (49). We demonstrated previously that LL-37 triggers LTB₄ release from human macrophages via P2X₇R (8), and it was reported that LTB₄ formation induced by ATP via P2X₇R was required for parasitic L. amazonensis elimination (50). All of this evidence suggests that P2X₇R-mediated LL-37 internalization is a novel mechanism by which P2X₇R contributes to pathogen clearance.

Internalized LL-37 could be degraded further by macrophages. LL-37 not only kills pathogens directly, it also is involved in immunomodulation and inflammatory reactions. Excessive expression of LL-37 was reported in many chronic inflammatory diseases (2). Previous studies showed that several enzymes, such as mast cell tryptase (51) and neutrophil elastase and cathepsin D (52), can further digest LL-37. It appears that LL-37 is engulfed first by macrophages to exert a bactericidal effect and then is eliminated by intracellular lysosomal degradation, possibly limiting the detrimental effects of cathelicidin during inflammation.

In conclusion, our study shows that human macrophages internalize antimicrobial peptide LL-37 via a P2X₇R-dependent mechanism primarily involving a clathrin-dependent pathway, which leads to improved pathogen clearance. Several lines of evidence suggest that LL-37 and other antimicrobial peptides are promising molecules for the treatment of chronically infected wounds and inflammatory diseases (53, 54). Hence, our findings provide new insights and strategies for the potential therapeutic uses of LL-37.

We thank Prof. Birgitta Agerberth for scientific discussions and for providing LL-37 Ab. We also thank the Center for Live Imaging of Cells at Karolinska Institutet for technical support with confocal microscopy.

The authors have no financial conflicts of interest.