|Home | About | Journals | Submit | Contact Us | Français|
α-Hemolysin (HlyA) from Escherichia coli and leukotoxin A (LtxA) from Aggregatibacter actinomycetemcomitans are important virulence factors in ascending urinary tract infections and aggressive periodontitis, respectively. The extracellular signaling molecule ATP is released immediately after insertion of the toxins into plasma membranes and, via P2X receptors, is essential for the erythrocyte damage inflicted by these toxins. Moreover, ATP signaling is required for the ensuing recognition and phagocytosis of damaged erythrocytes by the monocytic cell line THP-1. Here, we investigate how these toxins affect THP-1 monocyte function. We demonstrate that both toxins trigger early ATP release and a following increase in the intracellular Ca2+ concentration ([Ca2+]i) in THP-1 monocytes. The HlyA- and LtxA-induced [Ca2+]i response is diminished by the P2 receptor antagonist in a pattern that fits the functional P2 receptor expression in these cells. Both toxins are capable of lysing THP-1 cells, with LtxA being more aggressive. Either desensitization or blockage of P2X1, P2X4, or P2X7 receptors markedly reduces toxin-induced cytolysis. This pattern is paralleled in freshly isolated human monocytes from healthy volunteers. Interestingly, only a minor fraction of the toxin-damaged THP-1 monocytes eventually lyse. P2X7 receptor inhibition generally prevents cell damage, except from a distinct cell shrinkage that prevails in response to the toxins. Moreover, we find that preexposure to HlyA preserves the capacity of THP-1 monocytes to phagocytose damaged erythrocytes and may induce readiness to discriminate between damaged and healthy erythrocytes. These findings suggest a new pharmacological target for protecting monocytes during exposure to pore-forming cytolysins during infection or injury.
α-Hemolysin (HlyA) is an important virulence factor frequently produced by strains of pathogenic Escherichia coli (1,–3). The frequency with which HlyA-producing E. coli strains are isolated from patients increases with severity of the disease (for a review, see reference 2). HlyA is a pore-forming repeat in toxin (RTX) family member which inserts itself receptor independently into cell membranes (1). The cytotoxic effect of HlyA is massively amplified by ATP release, presumably through the HlyA pore (4) and following P2X receptor activation (5, 6). In erythrocytes, P2X1 and P2X7 receptors have been implicated in HlyA-induced hemolysis, and blocking of either of these receptors substantially reduces the hemolysis (5, 6). Interestingly, insertion of a HlyA pore does not cause immediate cell swelling and rupture but initially triggers a significant volume reduction that results from an increase in the intracellular Ca2+ concentration ([Ca2+]i) followed by activation of the Ca2+-sensitive K+ and Cl− channels KCa3.1 and TMEM16A (7). During erythrocyte shrinkage, cells expose phosphatidyl serine (PS) in the outer plasma membrane leaflet (7). Recently, we discovered that THP-1 monocytes are more likely to recognize and phagocytose erythrocytes that have been exposed to HlyA (8). This phagocytosis is prevented if HlyA-induced cell damage is diminished by P2X receptor antagonists or if cell shrinkage and PS exposure are blocked (8).
Leukotoxin A (LtxA) is a virulence factor often released from Aggregatibacter actinomycetemcomitans in the periodontal connective tissue (9,–11). The amount of LtxA released by a given substrain of A. actinomycetemcomitans varies, and the release of LtxA is particularly high from the JP2 substrain that lacks 530 bp of the leukotoxin promoter region (12). The JP2 variant has been found to associate with more aggressive forms of A. actinomycetemcomitans-induced gingivitis (9). Similar to HlyA, LtxA shows marked erythrotoxic effects via ATP release and subsequent P2X1 and P2X7 receptor activation (13). LtxA also belongs to the RTX family but is, in terms of cell lysis, more aggressive toward macrophages and leukocytes (14,–16). Interestingly, the lysis of human monocytes involves activation of caspase 1 and release of interleukin-1β (IL-1β) (17, 18), which is closely associated with P2X7 activation (for a review, see references 19 and 20). The selectivity for these cells is mediated via surface expression of β2-integrin (21), which is a transmembrane glycoprotein that is paired with a distinct α chain: αLβ2 (CD11a/CD18, leukocyte function-associated antigen 1 [LFA-1]), αMβ2 (CD11b/CD18, complement receptor 3, Mac-1), αXβ2 (CD11c/CD18, complement receptor 4, p150/195), and αDβ2 (CD11d/CD18) (22, 23). Apparently, the affiliation of LtxA with cell membranes relies on β2-integrin expression mainly and is independent of the associated α-integrin expression (24). LtxA, similar to HlyA, triggers erythrocyte shrinkage and PS exposure in the outer leaflet prior to cell lysis (13).
To monocytes, PS exposure is a strong incitement to phagocytose damaged erythrocytes (25, 26). Specifically, THP-1 monocytes are shown to phagocytose erythrocytes damaged by HlyA (8), and in vivo monocytes are likely to become exposed directly to the toxin during this process. Therefore, it is important to know how monocytes react to pore-forming virulence factors and to the concomitant ATP release. Here, we show that a variety of P2 receptors (P2X1, P2X4, and P2Y2) participate in the ATP-induced [Ca2+]i response in THP-1 monocytes. Both HlyA and LtxA cause acute release of ATP and an ensuing increase in [Ca2+]i, which in part is secondary to activation of the functionally expressed P2 receptors. The two RTXs are capable of causing lysis of both freshly isolated human monocytes and THP-1 monocytes in a P2X1, P2X4, and P2X7 receptor-dependent fashion. Notably, we find that a far larger population of THP-1 cells is affected by the toxin than the ones that actually lyse upon exposure to the toxins. These cells show marked uptake of propidium iodide. Inhibition of P2X7 receptors prevents the later stage of cell damage but does not prevent cell shrinkage inflicted by toxin exposure, whereas P2X1 receptor inhibition only prevents cell lysis. Interestingly, preconditioning the THP-1 cells with HlyA preserves the capacity of the monocytes to recognize and phagocytose damaged erythrocytes, whereas this was not the case with LtxA. Taken together, this study shows that P2X receptor inhibition clearly reduces the toxin-induced cell damage and that the phagocytic capacity of THP-1 monocytes is uncompromised by HlyA exposure.
Tamm-Horsefall protein 1 (THP-1), a human monocyte cell line from the American Type Culture Collection (Manassas, VA, USA), was grown in cell flasks in RPMI 1640 supplemented with 10% fetal bovine serum (Gibco, Grand Island, NY, USA) and antibiotics (penicillin at 1 U ml−1 and streptomycin at 100 μg ml−1). The cells were supplemented with medium every other day and passaged with 7-day intervals. Macrophages were isolated from BALB/c P2X7+/+ and P2X7−/− mice by peritoneal lavage. Anesthetized mice were injected with HEPES-buffered salt solution (HBS) into the peritoneal cavity, and mouse abdomens were closed with a pèan and massaged for 1 min. Peritoneal fluid was collected, centrifuged at 1,300 × g for 3 min, and resuspended in HBS. Macrophages were isolated by seeding on fibronectin-1-coated coverslips where the nonadherent cells were removed by washing. Human blood was collected by venopuncture from healthy volunteers and isolated as previously described (5) and according to permission given by the Danish Scientific Ethics Committee (M20110217). Human primary monocytes were isolated from patient buffy coats received from healthy volunteers from the blood bank and isolated by Ficoll centrifugation according to the manufacturer's instructions. Monocytes were isolated by continuous counterflow centrifugal elutriation using an Avanti J-26 XP (Beckman Coulter, Ramcon, Denmark), and fractions were collected. The purity of fractions was consolidated using a NovoCyte flow cytometer (ACEA Biosciences, AH Diagnostic, Denmark).
HlyA was purified from the supernatant from the E. coli strain ARD6 (serotype O6:K13:H1) grown in lysogeny broth (LB) medium in a process modified from the method described by Hyland et al. (83). A colony of E. coli ARD6 was transferred to 4 ml sterile LB medium and incubated overnight with constant swirling (37°C, 200 rpm). The following morning, 1 ml of the overnight culture was transferred to 1 liter sterile LB medium supplemented with 10 mM CaCl2 and incubated for 4.5 h in a shaker (37°C, 200 rpm). After incubation, cultures were centrifuged twice (2,943 × g, 15 min, 4°C) to pellet the bacteria. The supernatant was sterile filtered (pore size, 0.22 μm; Millipore, Bedford, Massachusetts) and pH adjusted to 4.5 (1 M malonic acid). HlyA was precipitated overnight with ethanol (25%, vol/vol, 4°C). The precipitate was centrifuged (17,300 × g, 30 min, 4°C; Sorvall RC-5C; Thermo Scientific), and the pellet was resuspended in 6 M guanidine-HCl, precipitated for 60 min with ethanol (90%, vol/vol, −20°C), and centrifuged at 12,960 × g. The final precipitate containing HlyA was resuspended in a Tris-buffered 8 M guanidine-HCl solution with 10 mM dithiothreitol (DTT) (pH 6.0). Purification of LtxA from A. actinomycetemcomitans strain HK921 (JP2 positive) was performed according to the methods described by Reinholdt et al. (24). LtxA was purified from bacterial supernatant using an ACTA 900 purifier system (GE Healthcare) and then eluted through a NaCl gradient and cation exchanger (Source S; GE Healthcare). The content of LtxA was analyzed by SDS-PAGE, and material containing LtxA was isolated and concentrated using a Centriprep YM 10 filtration unit (Millipore).
THP-1 monocytes were washed and suspended in HBS (0.5 × 106 cells ml−1) and incubated with calcein-AM (5 μM, 30 min, 37°C, 150 rpm). Cells were washed, resuspended, and divided among wells of a 96-well plate (final volume, 200 μl). The cells were incubated in the dark with compounds or vehicles and toxins (150 rpm, 60 min, 37°C). The experiment was terminated by centrifugation (1,609 × g, 3 min). Fluorescence of the top supernatant (100 μl) was determined by a fluorescence reader (excitation wavelength, 488 nm; emission wavelength, >520 nm; Mithras LB 940; Berthold Technologies, Bad Wildbad, Germany). Standard curves with increasing concentrations of toxin identified the concentration which caused 50% lysis of 1.25% (vol/vol) human erythrocytes during 60 min of incubation at 37°C (i.e., the 50% effective concentration [EC50]). Lytic activity of both HlyA and LtxA as well as THP-1 cell permeabilization were verified by light microscopy.
Washed THP-1 monocytes were centrifuged at 280 × g, suspended in HBS (0.5 × 106 cells ml−1), and incubated with a propidium iodide (PI) cell permeabilization kit (5 μM, 5 min, 37°C; Invitrogen). Cells were incubated in blockers of interest in Eppendorf tubes for 5 min before addition of either HlyA (EC50) or LtxA (EC50). The cells were incubated in the dark with toxins in the presence of compounds or vehicles (150 rpm, 60 min, 37°C). Each sample was assessed by flow cytometry (BD Accuri C6l BD Biosciences, Albertslund, Denmark). The plot was gated for monocytes (high forward scatter and high side scatter) and propidium iodide excited at 488 nm.
Peritoneal macrophages on fibronectin-1-coated coverslips were incubated with fluo-4-AM (5 μM, 60 min, 37°C). The preparations were placed in a cell chamber on an IMIC stage (TILL Photonics, Munich, Germany) and observed at 37°C. The fluorophore was excited at 488 nm, and emission was collected at >510 nm. The preparations were imaged with a 40×, 1.45-numeric-aperture Plan Apo (Olympus) objective and a charge-coupled-device camera (Sensicam qe; PCO, Kelheim, Germany) for live cell imaging. The entire setup was from Bio-Science ApS, Gilleleje, Denmark.
THP-1 monocytes in suspension were incubated with fluo-4-AM (5 μM, 30 min), washed, resuspended, and divided into transparent 96-well plates. Time-lapse experiments were carried for 30 min with 8 to 15 measurements per min depending on the number of wells included in the experiment. Fluorescence was recorded in a luminometer (Mithras LB 940; Berthold Technologies, Bad Wildbad, Germany) after excitation at 488 nm and emission at >510 nm.
Extracellular ATP was measured via firefly luciferase, which catalyzes the oxidation of luciferin in the presence of ATP and produces luminescence. The luminescence signal was recorded on a plate reader (Mithras LB 940; Berthold Technologies, Bad Wildbad, Germany). The THP-1 monocytes were incubated with RTX for 1, 2, 5, 10, 20, and 30 min and the supernatant was collected. The supernatant was heated (100°C, 30 s) to inhibit potential ecto-ATPase activity and frozen until measurements were conducted according to the manufacturer's protocol.
THP-1 cells were grown in collagen 1-coated 96-well plates (BD BioCoat collagen 1, rat tail; Thermo Fisher Scientific Inc., Waltham, MA, USA). Before the experiment the culture medium was removed from the wells. Erythrocytes loaded with calcein-AM were washed, subjected to toxin, HBS, or ionomycin, and then washed again. After this wash, the erythrocytes were resuspended to a concentration of 0.1% (vol/vol), added to the THP-1 cells in the 96-well plates, and incubated for 60 min at 37°C in HBS in the dark with constant swirling (150 rpm). After termination of incubation, the wells were washed twice in HBS. The washing procedure was tested in a series of pilot experiments to achieve wells with no remaining erythrocytes without removing the layer of THP-1 cells at the bottom of the well. The fluorescence of the wells was determined by excitation at 488 nm and emission at >520 nm in a fluorescence plate reader (Mithras LB 940).
Washed erythrocytes were suspended in HBS to produce an erythrocyte volume fraction of 1.25% in the final test solution. HlyA was added in increasing concentrations for 60 min at 37°C under constant swirling (180 rpm) in 96-well plates. The experiment was terminated by centrifugation of the samples at 2,325 × g for 3 min. The optical density of the supernatant was determined at 540 nm in a plate reader (PowerWave microplate spectrophotometer; Biotek Instruments, Winooski, VT, USA) as a measure of the hemolytic activity. Standard curves with increasing concentrations of HlyA identified the EC50 of the toxin. The EC50 for LtxA was found at a concentration of 0.122 μg ml−1, which is slightly more potent in this purification than what was previously published (13). The EC50 for HlyA varies and has to be determined every experimental week because of the faster degradation of protein activity. The amount of active protein which corresponded to EC50 in our preparation has been determined to be 25 ng ml−1 (4).
HBS consisted of the following (in mM) at pH 7.4 and 37°C: [Na+], 138.0; [Cl−], 132.9; [K+], 5.3; [Ca2+], 1.8; [Mg2+], 0.8; [SO42−], 0.8; [HEPES], 14; [glucose], 5.6. ATP-2′,3′-dialdehyde (oxidized ATP), ATP, UTP, NF449, 5-BDBD, A804598, EGTA, U73122, and ionomycin were obtained from Sigma-Aldrich. Brilliant blue G (BBG) was purchased from INC Biomedicals (Aurora, OH), 2-APB was from Tocris/Bio-techne (Bristol, United Kingdom), and fluo-4-AM, calcein-AM, and an ATP determination kit were from Invitrogen (Taastrup, Denmark). Fibronectin-coated coverslips were supplied by NeuVitro (Vancouver, WA, USA).
Data are presented as means ± standard errors of the means (SEM), and the n value indicates the number of experiments. The data were tested for normality by Kolmogorov-Smirnov test, and then the data were submitted to paired or unpaired Student's t test or one way analysis of variance (Tukey posttest) as appropriate.
In erythrocytes, there is a close association between the cytotoxic effects of HlyA and LtxA and P2X receptor activation (5, 13). Inhibition of P2X1 and P2X7 receptors completely abolishes HlyA- and LtxA-induced hemolysis (5, 13). HlyA and LtxA both cause PS exposure in the outer leaflet (7), which triggers phagocytosis of damaged erythrocytes (8). Membrane insertion of either HlyA or LtxA releases ATP during the erythrocytal shrinkage phase, well before any cell lysis is detected (4). Thus, the cells which are to be recognized and phagocytosed expose monocytes and macrophages to massive amounts of extracellular ATP. Therefore, it is important to know how ATP and the toxin itself interfere with the phagocytic process. In the case of LtxA, polymorphonuclear leukocytes (PMN) are prevented from engulfing bacteria because of the specific lytic actions of LtxA locally on leukocytes (27). This may explain the chronic gingival infection caused by the JP2 clone of A. actinomycetemcomitans. Regarding erythrocyte phagocytosis, it is important to know whether P2 receptor signaling potentiates the HlyA- or LtxA-induced phagocyte damage. As an initial step, we characterized the functional expression of P2 receptors in THP-1 cells as a reference for the toxin-induced effects.
THP-1 monocytes (28,–31), like their close relatives, monocytes (32,–35) and macrophages (36,–41), express a large variety of P2Y and P2X receptors. Figure 1 shows the concentration-response relationship for ATP and UTP in THP-1 monocytes. ATP clearly increases [Ca2+]i even at very low concentrations (10−8 M, P = 0.001). Notably, the concentration-response curve does not show a simple sigmoid relationship between the increase in [Ca2+]i and the concentration of extracellular ATP, indicated by the two EC50s in Fig. 1B. Rather, the initial response is to very low concentrations before there is a second rise at concentrations above 1 μM. This seemingly biphasic concentration-response relationship suggests that more types of P2 receptors with different sensitivities to ATP as agonist are functionally expressed in the cells and contribute to the collective ATP response. In contrast, the response to UTP, which selectively activates P2Y receptors, does not show multiple phases, which suggests activation of only one receptor subtype or that the receptor subtypes involved all have a very similar sensitivity to UTP. It is noteworthy that the cells require notably higher concentrations of UTP to show a detectable response and that the EC50 for UTP is around 5 μM, whereas for ATP it is approximately 1 μM lower. This difference is likely caused by expression of the highly ATP-sensitive P2X1 receptors rather than a relatively low-level expression of P2Y receptors compared to P2X receptors. The reason is that the maximal amplitude of the UTP-induced [Ca2+]i is not statistically significantly different from the response inflicted by ATP, which argues for a substantial functional amount of P2Y receptors.
In terms of P2X receptors, the P2X7 receptor is the best-characterized P2 receptor in THP-1 monocytes (28, 42). Surprisingly, brilliant blue G (BBG), which shows some selectivity for P2X7R, had no effect on either the amplitude of the peak or on the plateau phase of the ATP-induced [Ca2+]i response in THP-1 cells (Fig. 2A). This is consistent with findings from Ase et al., who similarly were unable to detect an effect of BBG on the ATP-induced [Ca2+]i response in THP-1 cells (29). This suggests that P2X7 receptors do not have any quantitative relevance for the magnitude of the ATP-induced [Ca2+]i response in THP-1 monocytes. To confirm this, another antagonist with selectivity for the P2X7 receptor, oxATP, was applied at both higher and lower ATP concentrations. Even more surprising, oxATP resulted in higher amplitude of the ATP-induced [Ca2+]i response (Fig. 2B), which likely results from the ability of oxATP to inhibit ectoATPases and to reduce the extracellular degradation of ATP (43). The result does, however, support that P2X7 does not contribute significantly to the ATP-induced [Ca2+]i response. We wondered whether this lack of participation of P2X7 in the collected ATP-induced [Ca2+]i response was a general feature for monocytes/macrophages; thus, we isolated peritoneal macrophages from P2X7+/+ and P2X7−/− mice. In P2X7+/+ cells, ATP (10 μM) application caused a rapid oscillatory rise in [Ca2+]i followed by a slightly attenuated response to ATP (1 mM) which did not show the prolonged oscillation observed after the first stimulation. In comparison, the initial response to ATP (10 μM) was decreased in macrophages from P2X7−/− mice, whereas the response to a second exposure to ATP (1 mM) was identical to the first response. Apparently, P2X7 shows a more prominent function in murine macrophages than in human THP-1 cells, consistent with the P2X7 receptor expression levels increasing upon monocyte differentiation to macrophages (44). In monocytes, P2X7 has, based on mRNA levels, been proposed to be more abundant than P2X4 (32), and since the P2X4 receptor also is expressed in THP-1 monocytes (29), we tested the effect of the antagonist 5-BDBD with selectivity for P2X4 receptors (Fig. 2E). This antagonist (100 μM) clearly reduced the amplitude of the ATP-induced [Ca2+]i response, indicating that the P2X4 receptor is functionally expressed in the THP-1 monocytes.
As previously noted, THP-1 monocytes have a very high sensitivity to ATP, which is likely to be caused by P2X1 receptors. P2X1 receptors are known to be functionally expressed in murine macrophages (37, 38) but never, to our knowledge, in THP-1 monocytes. Figure 2D shows that the [Ca2+]i response caused by very low ATP concentrations (0.1 μM and 1 μM) were inhibited by the P2X1 antagonist NF449 (100 μM). This argues strongly that the P2X1 receptor is partially responsible for the collected ATP-induced response in THP-1 monocytes. In conclusion, the THP-1 monocytes show expression of P2X1 and P2X4 with [Ca2+] as a functional read out. There is, however, potentially room for other P2 receptors to participate in the ATP-induced response.
We addressed functional expression of P2Y receptors in these cells. Since antagonists for P2Y2 receptors are not very selective, we instead blocked the signal transduction pathways downstream of the P2Y receptors to assess their contribution to the overall ATP response. Since P2Y2 receptors previously shown to be expressed in THP-1 monocytes on the mRNA level (30, 45) show equipotent responses to ATP and UTP (46), we used UTP to selectively activate P2Y receptors without interfering with P2X receptors. It must be stressed that the pharmacological approach is the most favorable way of addressing P2Y2 protein expression because of the quality of the antibodies. Since several of the P2Y receptors shown to be expressed in the THP-1 cells by mRNA are Gxq coupled, we tested whether interference with phospholipase C (PLC) activation and the inositol trisphosphate (IP3) receptor would affect the UTP-induced [Ca2+]i response. Figure S1 in the supplemental material shows a reduced UTP-induced [Ca2+]i response by blockage of PLC (U73122) and IP3 (2-APB). Interestingly, blockage of PLC caused a minor but statistically significant reduction of the ATP-induced [Ca2+]i response, whereas this was not the case for blocking of the IP3 receptor (see Fig. S1C and D). However, removal of extracellular Ca2+ immediately before addition of ATP completely abolished the ATP-induced [Ca2+]i response. Collectively these findings indicate that the majority of the ATP-induced [Ca2+]i response indeed is mediated primarily through P2X receptors in THP-1 monocytes.
Since HlyA and LtxA have been demonstrated to release ATP immediately upon insertion into plasma membranes and THP-1 monocytes express several functional P2 receptors, we expect HlyA to trigger a [Ca2+]i response in THP-1 monocytes. Therefore, we exposed fluo-4-AM-loaded THP-1 monocytes to HlyA, which concentration-dependently increased [Ca2+]i (Fig. 3A and andB).B). The kinetics of the HlyA-induced [Ca2+]i response was much slower than that inflicted by ATP (Fig. 3C). However, similar to the ATP-induced [Ca2+]i response, the HlyA-induced [Ca2+]i increase was abolished by removing extracellular Ca2+ immediately before application of HlyA (Fig. 3D). This indicates that the HlyA-induced [Ca2+]i response is caused by Ca2+ entry either through the pore or via subsequent stimulation of Ca2+-permeable channels.
To determine whether ATP was indeed released in the early phase after addition of the toxins, in parallel we measured ATP release, [Ca2+]i, and cell lysis. Both HlyA and LtxA showed increments even at the earliest time points, whereas the maximal [Ca2+]i levels are found approximately 100 s later (Fig. 4A to toD).D). Notably, detectable cell lysis is first measured over 20 min later than these early events (Fig. 4E). These data support our previous results from erythrocytes and ATP-containing phospholipid vesicles (4) and underlines ATP release as one of the early events in RTX-induced cell damage.
In order to unravel if P2X receptors participated in the HlyA-induced [Ca2+]i response, we applied oxATP or BBG, which both show selectivity toward the P2X7 receptor. Similar to the ATP-induced [Ca2+]i response, both substances clearly enhanced the HlyA-induced [Ca2+]i response (Fig. 4A and andB).B). However, blocking of the P2X1 receptor with NF449 resulted in a marked reduction of the amplitude of the HlyA-induced [Ca2+]i response (Fig. 5D), whereas blocking P2X4 receptors by 5-BDBD did not show any effect. Interestingly, blockage of PLC caused a prolonged and intensified HlyA-induced [Ca2+]i response that did not return to baseline, whereas blocking the IP3 receptors alone by 2-APB had little or no effect on the fluorescence maximum (Fig. 5E). In this context it is interesting that phosphatidylinositol 4,5-bisphosphate (PIP2) has been shown to intensify the activity of P2X1, P2X4, and P2X7 receptors (47,–49), which may explain the potentiated response when PLC is inhibited. Note that the remaining HlyA-induced [Ca2+]i response was abolished by removal of extracellular Ca2+ (see Fig. S2 in the supplemental material).
LtxA is an RTX toxin, and it is still debated whether it forms actual pores in biological membranes (50,–53). Peculiarly, we could demonstrate that LtxA releases ATP from phospholipid-1-palmitoyl-2-oleoyl-phosphatidyl choline (POPC) vesicles with kinetics quite similar to those of HlyA. This is immediately consistent with ATP being released through a toxin pore. The finding could, however, reflect equally well that ATP slips past the membrane during the process of toxin insertion or that the general ATP leak increases from the cell as a consequence of membrane instability provided by LtxA. Since we know that LtxA causes nonlytic ATP release both from erythrocytes, vesicles, and now also THP-1 cells, it is not surprising that the toxin changes the [Ca2+]i when it interacts with THP-1 monocytes.
The [Ca2+]i response inflicted by LtxA shows a different pattern than that of HlyA. LtxA caused a relatively rapid but small peak in [Ca2+]i followed by a plateau that gained a higher amplitude than the original [Ca2+]i peak (Fig. 6A and andB).B). As the concentration of LtxA increases, the initial peak became less apparent. The initial [Ca2+]i peak was prevented by blockage of P2X1 receptors (Fig. 6C and andD),D), whereas the amplitude of the plateau was markedly reduced by specific blockage of P2X7 by A804598 and unaffected by P2X4 blockage (Fig. 6E). These results show that both P2X1 and P2X7 are implicated in the LtxA-induced [Ca2+]i response. Intriguingly, the data support a much larger contribution of P2X receptors in the LtxA-induced [Ca2+]i response compared to that of HlyA, which may reflect the relative permeability to Ca2+ through the LtxA protein, which itself is substantially lower than that through HlyA. This may support the notion that LtxA does not form stable pores in the plasma membrane; thus, Ca2+ entry into the cells will occur via the ATP-activated P2X receptors.
During phagocytosis of cytolysin-damaged erythrocytes, monocytes and macrophages are prone to be exposed to the given toxins. Notably, HlyA and LtxA both cause concentration-dependent lysis of THP-1 monocytes (Fig. 7A and andC).C). As expected, LtxA was much more potent at inducing THP-1 monocyte lysis, whereas HlyA (at the EC50 for human erythrocytes) only lyses ~10% of the THP-1 monocytes. Blockage of P2X1 (NF449), P2X4 (5-BDBD), and P2X7 receptor (A804598 and oxATP) all reduced lysis of HlyA-exposed THP-1 monocytes (Fig. 7B). Similar to HlyA, P2X receptors are also essential for the LtxA-induced cell lysis. Here, antagonists with selectivity for P2X1, P2X4, and P2X7 reduced the LtxA-induced lysis of THP-1 monocytes (Fig. 7D). In analyzing the inhibition pattern of P2X receptor blockers and the efficiency by which they reduce cell lysis, we conclude that an initial increase in [Ca2+]i, just like in erythrocytes, is unrelated to the following cell lysis. An alternative method to address the role of P2 receptor in a given response is to desensitize them with preexposure to ATP. Thus, we preincubated THP-1 monocytes with ATP, and the effects on HlyA- and LtxA-induced cell lysis are shown in Fig. S3 in the supplemental material. Preincubation with ATP reduced the HlyA-induced cell lysis at low concentrations of extracellular ATP, whereas the LtxA-induced lysis is massively reduced by preincubation with high concentrations of ATP. Combining the effect of increasing ATP concentrations with inhibition of P2X receptors revealed that a combination of A804598 and the P2X1 blocker NF449 completely abolished HlyA-induced cell lysis at all concentrations of ATP (see Fig. S4A). Similarly, blockage of P2X7 alone reduces the LtxA-induced lysis in THP-1 monocytes at all ATP concentrations (see Fig. S4B). This suggests that human monocytes have a built-in resistance mechanism when migrating into ATP-leaking tissue that may result from redistribution of P2X7 receptors, which occurs upon prolonged agonist stimulation in THP-1 monocytes (42). In agreement with the data on cultured monocytes, both HlyA and LtxA cause lysis of freshly isolated human monocytes from healthy volunteers. Inhibition of P2X1 and P2X7 markedly reduced lysis of human monocytes after exposure to HlyA (Fig. 7E). The effects are more modest when the lysis was caused by LtxA but nevertheless was of potential clinical relevance given the role of monocytes during the rapid innate immune response during the early stages of infection.
Since ATP is apparently released from THP-1 monocytes upon exposure to HlyA, we also addressed whether there is a role of P2Y receptors in HlyA-induced lysis of THP-1 monocytes. Figure 8 shows a concentration-dependent potentiation of UTP in HlyA-induced THP-1 cell lysis. Thus, activation of P2Y receptors can potentiate the ATP-induced amplification of HlyA-induced cytotoxicity in THP-1 monocytes even further.
Figure 9 illustrates that HlyA or LtxA causes calcein-loaded cells to leak the fluorophore into the surroundings without actual disintegration of the cells. The arrows in both Fig. 9A and andBB point out cells were the cellular fluorescence clearly reduces without notable disintegration of the cells. This cellular leakage is, similar to the lysis, prevented by P2X7 receptor inhibition.
ATP is known to cause cell damage by permeabilizing cells in a P2X7 receptor-dependent fashion and to allow propidium derivatives to pass within seconds of stimulation (54). Thus, a proportion of THP-1 monocytes which do not lyse may be damaged and go into apoptosis. To increase the robustness of our results, we assessed the toxic activity of HlyA and LtxA using propidium iodide as an indicator of cell permeabilization. The degree of intracellular fluorescence is a result of the extent and/or time by which the plasma membrane has been permeabilized during the experiment. This gave us the option to discriminate between healthy cells and dead cells and also to identify an intermediate low-fluorescence fraction of injured but still living cells. Clearly a much larger fraction of cells showed propidium iodide uptake compared to the fraction that lyses. Again, we saw a protective role of blockage of P2X1 and P2X7 for the cell damage inflicted by both toxins (Fig. 10B and andD).D). It is quite apparent that blockage of P2X7 reduced the number of injured cells, whereas this was not the case for P2X1. In the presence of the P2X7 receptor antagonist A804598, cell shrinkage, seen as a fall in forward scatter, is observed when cells are exposed to either HlyA or LtxA. However, in the presence of P2X7 receptor antagonist the cells do not progress onward toward permeation and lysis, consistent with the inhibition of a major conductance pathway in the cells important for final cell permeability. This finding is consistent with the pore dilation characteristics of the P2X7 receptor caused by prolonged activation by ATP and the finding that this pore allows inflow of propidium iodide. These data also confirm our previous findings that the P2X receptor inhibitors do not prevent the toxin from inserting into the membrane.
Attack by cytolysins is likely to affect the THP-1 monocytes' ability to phagocytose damaged erythrocytes. Surprisingly, we found that despite preexposure of THP-1 monocytes to HlyA (EC25), the full phagocytic function was preserved (Fig. 11C). Figure 11 illustrates that the phagocytosis of erythrocytes exposed to either ionomycin (1 μM) or HlyA (EC25) for 10 min is stronger when the THP-1 monocytes have been preexposed to HlyA (EC25). The same was observed for HlyA at EC50 (data not shown). This picture was different for LtxA, where preincubation with LtxA (EC25) merely reduced the phagocytic capacity for ionomycin-treated erythrocytes (Fig. 11D), most probably due to the high toxicity toward the monocytes. These results inform us that macrophages are relatively protected against HlyA and can maintain substantial function even in the continuous presence of the toxin. Inhibition of P2X will provide further protection of the cells against toxin-induced cell damage. In contrast, LtxA causes a high level of cytotoxicity, which is likely linked to its success as a virulence factor for A. actinomycetemcomitans. P2X receptor inhibition provides a distinct protection against this and may detrimentally affect the cause of a chronic infection with the bacterium.
Purinergic signaling is an essential component for normal monocyte/macrophage function (36, 41, 55, 56); thus, it plays a key role in regulating several immunological functions. Monocytes and macrophages are known to express several P2 receptors, of which the P2X1, P2X4, P2X7, and P2Y2 receptors are predominant (32, 38, 42). These receptors regulate key monocyte/macrophage functions, such as migration (56, 57), and the expression pattern of P2 receptors changes upon differentiation to tissue macrophages (31, 58). One of the essential functions of monocytes and macrophages is to clear damaged and/or senescent erythrocytes from the circulation (59). Erythrocytes are severely damaged by HlyA and LtxA (5, 13), which both initially cause severe cell shrinkage and PS exposure in a protracted process before they finally swell and lyse (7, 13). Because free hemoglobin has detrimental implications for the prognosis of, for example, sepsis (60, 61), it is important that the damaged erythrocytes are recognized and phagocytosed before intravascular lysis occurs. In this context, it is important that THP-1 monocytes are able to recognize and phagocytose erythrocytes exposed to HlyA (8).
However, phagocytes are likely to work in a hostile milieu while clearing erythrocytes during infection with RTX-producing bacteria. In addition to the RTX toxins themselves, they will be exposed to ATP, potentially at high local concentrations which, because of the expression of P2X7 receptors, can result in acute membrane permeation of the phagocytic cells. Therefore, we were interested in how phagocytic cells cope with this type of environment. Here, we show that both P2 receptor subfamilies are functionally expressed in THP-1 monocytes. Our study confirms the presence of P2Y receptors with functional properties consistent with the previously demonstrated P2Y2 receptor (31) and P2X receptors that are likely to be P2X4 and P2X1. Surprisingly, the P2X7 receptor does not contribute to the ATP-induced [Ca2+]i response in THP-1 cells, although this was clearly seen in murine peritoneal macrophages. These data are consistent with previous data presented by Ase et al. (29) and with the notion that P2X7 receptor expression is more pronounced in mature macrophages (31, 58).
Both of the toxins studied here are known to release ATP after insertion into biological membranes (4). This release is nonlytic and occurs without activation of any of the known ATP release pathways. Since ATP release is seen immediately upon insertion of the toxins in artificial lipid vesicles without any detectable vesicular lysis, it was concluded that ATP may be released through the toxin pore. In line with this, both HlyA and LtxA cause early elevations in extracellular ATP with quite distinct kinetics, where the HlyA-induced ATP release is more transient than the LtxA-induced release. In any case, ATP release precedes both the maximal [Ca2+]i response to the toxin as well as cell lysis, which is first detectable over 20 min after the initial release of ATP. In support of ATP being an essential part of the initial response to the toxin, the toxin-induced [Ca2+]i increase was found to be at least partially secondary to P2X receptor activation. Similar to the ATP-induced [Ca2+]i response, we found that P2X7 receptors did not participate in the HlyA-induced [Ca2+]i response, whereas P2X1 and P2X4 did. There was no significant contribution from P2Y receptors to the HlyA-induced [Ca2+]i response, since neither PLC nor IP3 receptor blockage influenced this response. With regard to the LtxA-induced [Ca2+]i response, both the P2X1 and the P2X7 receptors seem to be involved. The P2X1 receptor is activated immediately upon toxin insertion, whereas the P2X7 receptor participated in the sustained, late response. Most of the LtxA-induced [Ca2+]i response is prevented by P2X receptor antagonists, which suggests that, unlike HlyA, LtxA does not in itself inflict a marked Ca2+ permeability. On this note, it is worth mentioning the controversy regarding whether LtxA forms an actual pore (50,–53, 62, 63), which is not yet fully settled. Our previous data suggest that ATP is released upon LtxA insertion into the membrane, and even though the most straightforward solution would be that ATP passes through a pore formed by the protein, we cannot exclude unselective ATP leak by membrane destabilization (4). Thus, it is still possible that the low Ca2+ permeability results from a lack of genuine pore formation by LtxA in the membrane.
In line with this, it may be surprising that in a previous study we were able to detect an effect of LtxA directly in artificial, plain lipid vesicles (4). As previously mentioned, β2-integrin is the membrane receptor that provides the selectivity by which LtxA targets leukocytes (14,–16). Despite this, LtxA is able to lyse cells that do not express β2 integrin, such as erythrocytes, from various species (13, 64). We have demonstrated that glucans terminated with a terminal sialic acid essential for the sensitivity of LtxA regardless of whether it is β2 integrin or another protein which bears the sialic acid residue (64). Removal of the sialic acid markedly reduces the negative surface charge in erythrocytes, and we found a correlation between the surface charge and the association of LtxA with the membrane (64). This means that LtxA can be associated with membranes by less specific interaction with the surface. Thus, the negative charge on the artificial phospholipids apparently is sufficient to allow membrane integration of LtxA, an interaction which was demonstrated by a combination of circular dichroism and fluorescence spectroscopy in addition to the distinct LtxA-induced, nonlytic ATP release from the vesicles.
It is worth noting that the kinetic of the toxin-induced [Ca2+]i response is distinctly different after application of HlyA compared to that after application of ATP. The slower rise in [Ca2+]i after exposure to HlyA is immediate, consistent with HlyA-induced [Ca2+]i change being a compound response, which in part results from Ca2+ influx through the pore itself. Moreover, the slower kinetics are likely to reflect that insertion of toxins in the membrane is a stochastic event.
In terms of lysis, the THP monocytes have a very low sensitivity to HlyA compared to the erythrocytes, where a concentration of HlyA which causes 50% lysis in human erythrocytes barely lyses 10% of the THP-1 monocytes. The THP-1 monocytes express β2-integrin (65, 66) and accordingly show a high sensitivity to LtxA. P2X receptor inhibition has been extraordinarily effective in reducing the lysis of erythrocytes not only to HlyA (5) and LtxA (13) but also to other types of bacterial toxins (alpha-toxin from Staphylococcus aureus , ApxIA hemolysin of Actinobacillus pleuropneumoniae , and β-toxin from Clostridium perfringens ) and complement activated membrane attack complexes (70). However, in tissues where P2Y receptors predominate, for instance, in renal epithelial cells, the effect of HlyA shows a completely different pattern (71, 72). In renal epithelia, HlyA does not cause cell lysis but rather triggers [Ca2+]i oscillations which result in release of interleukin-6 (71, 72), which both require P2Y2 receptor activation, whereas P2X receptors had no effect on renal epithelia (71). Thus, the effect of HlyA on a given tissue is largely determined by the expression pattern of P2 receptors and complies with the notion that tissue responses to HlyA require ATP signaling.
Since macrophages express both P2Y and P2X receptors, one would expect that either would be activated upon addition of the bacterial toxins and potentially modify cell lysis. Strikingly, the P2X7 receptor, which did not contribute to either the ATP- or HlyA-induced [Ca2+]i response, markedly reduced the THP-1 monocyte lysis of both HlyA and LtxA. Similar to erythrocytes, this effect is not exclusive to P2X7, since a reduction of the cytolysis is also observed after inhibition of P2X1 and P2X4 receptors. Our data confirm previous studies which demonstrated that oxATP reduced the LtxA-induced cell lysis in human leukocytes and macrophages (73) and underscores at least partially a P2X7-mediated effect. This pattern is reproduced in freshly isolated human monocytes, where inhibition of P2X1 or P2X7 markedly reduces HlyA-induced cell lysis. These findings fit the properties of the P2X7 receptor. Because of its apparent ability to dilate to a higher conductance state upon prolonged stimulation (42), it could contribute markedly to the overall conduction despite a low expression level and a modest contribution in brief acute ATP applications at lower concentrations. On this note, P2X7R activation has been closely associated with IL-1β release in macrophages. The conventional view is that ATP causes massive K+ loss through the P2X7 receptor that leads to activation of NOD-like receptor family, pyrin domain containing 3 (NRPD3) inflammasome activation, caspase 1 activation, and subsequent IL-1β release (for a review, see references 19 and 20). Interestingly, this pathway is essential for both LtxA- and HlyA-induced caspase 1 activation and IL-1β release from macrophages (17, 18, 74), underscoring the importance of P2X receptors in the cellular effects of the two toxins. Moreover, we found that desensitization of P2X receptors by preincubation with ATP at high concentrations reduced LtxA-induced lysis, whereas HlyA-induced lysis declined marginally at low concentrations. This is potentially pathophysiologically relevant, because LtxA is an important virulence factor during A. actinomycetemcomitans-induced gingivitis and the macrophages works in confined connecting tissue compartments, where the extracellular ATP concentrations may become high during infection. However, during urosepsis, HlyA acts either in the renal tubules or in the bloodstream where the cells have high abundance of ectoATPases (31, 75,–78); thus, monocytes/macrophages are likely to be exposed to low ATP concentrations only. In this connection, it is interesting that we found HlyA and LtxA damaged a far larger population of the THP-1 monocytes than those that are lysed by the cytolysins. Interestingly, the toxins in themselves lead to cell shrinkage of the THP-1 monocytes, and the P2X7 receptor is necessary to cause the swelling and disintegration by activations of pores large enough to let propidium iodide through. These pores are likely a direct dilation of the P2X7 receptor itself. The P2X1 receptor also contributes to the overall permeability of the THP monocytes; however, inhibition of the P2X1 receptor limits the cell damage and ultimately prevents cell death.
The THP-1 monocytes also express P2Y2 receptors, which is likely responsible for the HlyA-induced IP3 degradation observed in leukocytes (79). Here, we show that stimulation of P2Y receptors by UTP significantly potentiated cell lysis in THP-1 monocytes. This effect is not very likely to result from modulation of the P2XR function, since PIP2 is known to stimulate the activity of P2X1, P2X4, and P2X7 receptors (47,–49). Interestingly, P2Y receptor activation has previously been implicated in cytolysis. In erythrocytes infected with Plasmodium falciparum, P2Y1 receptor blockers significantly reduced hemolysis, and the hemolysis was significantly blunted in P2Y1−/− mice infected with P. berghei (80). These data underscore that ATP release and P2 receptors in general amplify the cytotoxic effects of bacterial cytolysins.
Because of the cytolytic effects, we expected the toxins to reduce the phagocytic capacity of THP-1 monocytes as previously demonstrated for PMN-dependent phagocytosis of A. actinomycetemcomitans (27). Surprisingly, preexposure of the THP-1 monocytes to small nonlytic but also lytic concentrations of HlyA did not reduce erythrophagocytosis but showed a slight tendency toward enhancing it. This toxin preconditioning may create readiness for the nonlysed THP-1 monocytes to recognize erythrocytes damaged by HlyA, since an increase in [Ca2+]i is known to support the phagocytic function (81, 82). This effect was not clear after preexposure to LtxA, presumably because of the larger fraction of lysed phagocytes.
In conclusion, ATP-induced P2X receptor activation is central to the cytotoxic effects of HlyA or LtxA on THP-1 monocytes. Inhibition of P2X receptors is essential to protect macrophages and monocytes and to preserve the overall phagocytic function under attack by cytolysin-producing bacteria. Thus, P2X receptor desensitization or inhibition may affect the overall prognosis under critical conditions with massive erythrocyte damage. Moreover, P2X receptor inhibition may provide a future pharmacological target for immune protection by improving monocyte survival during the initial stage of infection and thereby tilt the scales during immunological combat.
We thank Helle Jakobsen for her skilled laboratory expertise and her kind contribution to current data acquisition.
The study was funded by The Danish Council for Independent Research (DFF-1331-00203A and DFF-0602-02145B) and Lundbeck Foundation (R100-A9482).
Authors made the following contributions: S.K.F., idea generation, study design, performing of experiments, figure design, and manuscript preparation and finalizing; M.R.J., study design and finalizing the manuscript; M.S., idea generation, performing of experiments, and manuscript preparation; H.A.P., idea generation, study design, figure design, and manuscript preparation and finalizing.
None of the authors has any conflict of interest to declare in connection to this paper, and funding partners had no influence on any part of the scientific process.
Supplemental material for this article may be found at http://dx.doi.org/10.1128/IAI.00674-16.