Group IIA phospholipase A2 (PLA2-IIA) is a newly recognized antibacterial acute phase protein. The concentration of PLA2-IIA increases up to 500-fold in the blood plasma of patients with severe acute diseases, compared with healthy persons. Despite numerous studies, the exact roles of this enzyme in human diseases are unknown. This study investigated the antibacterial properties of PLA2-IIA in human acute phase serum. PLA2-IIA in serum samples of patients with bacterial infections was capable of killing 90% of Staphylococcus aureus and 99% of Listeria monocytogenes in vitro after incubation for 2 h. At concentrations found in normal human serum, PLA2-IIA killed 90% of L. monocytogenes but did not kill S. aureus or Escherichia coli. The bactericidal effects of acute phase and normal human serum were abolished after depletion of PLA2-IIA by immunoadsorption.
The concentration of group IIA phospholipase A2 (PLA2-IIA) increases markedly in blood plasma of patients with severe acute diseases such as sepsis [1, 2], peritonitis , and bacterial infections . The exact role(s) of PLA2-IIA in human diseases is unknown . The potential sources of PLA2-IIA found in human circulation include liver cells (hepatocytes) [5, 6] and blood eosinophils . This enzyme, even at very low concentrations, is capable of killing gram-positive bacteria in vitro . We demonstrated recently that overexpression of PLA2-IIA in transgenic mice protects these animals against experimental bacterial infections [9, 10]. To elucidate the role of PLA2-IIA in human diseases, we measured its antibacterial properties in serum of patients with bacterial infections.
Materials and Methods
Patients and serum samples. Serum was obtained from 8 patients with acute bacterial infections with high concentrations of PLA2-IIA (> 200 μg/L) in serum at arrival at the Turku University Hospital (Turku, Finland) outpatient unit. Serum samples of 6 patients with no infectious diseases and low concentrations of PLA2-IIA (< 3 μg/L) in serum served as control samples. The patients had not been treated with antibiotics within 3 weeks before serum sampling. Table 1 shows patient diagnoses.
Time-resolved fluoroimmunoassay (TR-FIA) for PLA2-IIA in serum. We measured the concentration of PLA2-IIA in serum by TR-FIA, as described elsewhere . In brief, serum samples were incubated in the wells of 96-well plates coated with protein A-purified polyclonal rabbit anti-human recombinant PLA2-IIA IgG and saturated with bovine serum albumin (BSA). After incubation for 1 h at 25°C with shaking, the plate was washed, and the wells were incubated with anti-human recombinant PLA2-IIA IgG labeled with Europium (Eu-labeling kit; Wallac). Fluorescence was measured with a 1230 Arcus fluorometer (Wallac). Human recombinant PLA2-IIA was used as a standard.
Bactericidal activity in serum. To investigate the antibacterial properties of circulating human PLA2-IIA, the bactericidal activity of serum against Staphylococcus aureus (ATCC 25923), Listeria monocytogenes (Finnish National Health Institute control strain), and Escherichia coli (ATCC 25922) was measured, as described elsewhere . About 5 × 105 bacteria cells in 20 μL of HEPES buffer (pH 7.4, 2.0 mmol/L Ca2+; pH 7.4, 10 mg/mL BSA) were incubated for 2 h at 37°C with 20 μL of acute phase serum or serum from healthy subjects. PLA2-IIA was removed from the serum by immunoadsorption, as described below. Samples were taken from the reaction mixtures at 30, 60, and 120 min, and colony-forming units were measured from logarithmic dilutions.
Adsorption of PLA2-IIA from acute-phase serum. A polyclonal rabbit antibody produced against recombinant human PLA2-IIA  was used to remove PLA2-IIA from serum by immunoadsorption, as described elsewhere , with slight modifications. In brief, 96-well plates were coated with protein A-purified rabbit anti-human PLA2-IIA IgG or corresponding IgG from preimmune rabbit serum. Serum samples were added to the anti-PLA2-IIA IgG-coated or preimmune IgG-coated wells, shaken for 30 min at 4°C, and transferred to another IgG-coated well. Incubations were repeated in 4 consecutive wells. The binding of PLA2-IIA to the wells was confirmed by measuring the time-resolved fluorescence in the wells with Eu-labeled anti-PLA2-IIA IgG, as described elsewhere . After immunoadsorption, the residual concentration of PLA2-IIA in serum was measured as described above. To obtain nondepleted serum rich in PLA2-IIA, the procedure was repeated with corresponding acute phase serum samples incubated in preimmune rabbit IgG-coated wells. Control serum samples of healthy persons (normal human serum [NHS] with low PLA2-IIA content; n = 6) were treated in the same way.
The bactericidal properties of PLA2-IIA in human acute phase serum against S. aureus and L. monocytogenes were confirmed by using acute phase serum immunoadsorbed with mouse monoclonal anti-human PLA2-IIA affinity sorbent, according to the manufacturer's instructions (Cayman Chemicals). We added 25 μL of immunosorbent into 250 μL of acute phase serum (patients 1 and 2; table 1), which was shaken at 1000 rpm for 60 min at 25°C and centrifuged at 500 g at 25°C for 15 min to deplete PLA2-IIA. About 99% of PLA2-IIA was removed by the immunosorbent. To return PLA2-IIA to the depleted serum samples, we added 3 μL of immunoaffinity-purified PLA2-IIA from human small intestine to 17 μL of depleted serum.
For immunoaffinity purification, polyclonal anti-human PLA2-IIA IgG  was coupled to NHS-activated HiTrap column (Amersham-Pharmacia), according to the manufacturer's instructions. To confirm that the restored bactericidal activity was due to the added small intestine PLA2-IIA, 3 μL of small intestine extract from which PLA2-IIA was depleted with the monoclonal anti-PLA2-IIA IgG affinity sorbent (Cayman Chemicals) was added to another 17-μL sample of PLA2-IIA-depleted serum (patients 1 and 2; table 1). Nontreated serum samples of the same patients served as positive (PLA2-IIA-rich serum) controls for immunoadsorption. Bactericidity was tested for all samples as described above.
The bactericidity test was performed against L. monocytogenes, also by using 2 serum samples from healthy persons (subjects 1 and 2; table 1). Depletion was done by the PLA2-IIA affinity sorbent in the same manner as with acute phase serum, as described above. To return PLA2-IIA to the depleted serum samples, 19 μL of depleted serum was mixed with 1 μL of small intestine PLA2-IIA diluted 20-fold with sterile saline. To confirm that the restored bactericidal activity was due to the added small intestine PLA2-IIA, 1-μL aliquots of 20-fold diluted small intestine PLA2-IIA extract depleted with monoclonal anti-PLA2-IIA IgG affinity sorbent were added to other 19-μL samples of PLA2-IIA-depleted serum (healthy subjects 1 and 2; table 1). Nontreated serum samples from the same subjects served as control samples.
To further study the role of PLA2-IIA in killing of S. aureus, 3 μL of small intestine PLA2-IIA was added to 17 μL of NHS, to increase the PLA2-IIA concentration to correspond to that of acute phase serum. To confirm that the restored bactericidal activity was due to the added small intestine PLA2-IIA, 3 μL of small intestine extract, depleted with the monoclonal anti-PLA2-IIA IgG affinity sorbent, were added to other 17-μL serum samples (healthy subjects 1 and 2; table 1). Nontreated serum samples from the same subjects served as control subjects. The bactericidity tests for all samples were done as described above.
To rule out the possible effects of shaking and centrifuging on the bactericidal properties of the serum samples, the bactericidity tests were repeated by using acute phase serum samples without affinity sorbent shaken at 1000 rpm for 60 min and centrifuged for 15 min at 500 g.
Testing of cross-reactivity of the polyclonal PLA2-IIA antibody. Because of new knowledge of different groups of secretory PLA2 subgroups [13, 14], TR-FIA was performed to test the polyclonal anti-human recombinant PLA2-IIA IgG used in the TR-FIA and depletion of PLA2-IIA for cross-reactivity against various other human recombinant secretory PLA2 subgroups. The PLA2s tested were of human groups IB, IIA, IIE, IIF, V, X, and XII (PLA2-IB, -IIA, -IIE, -IIF, -V, -X, and -XII, respectively) expressed in E. coli . The enzymes were gifts of Michael Gelb (University of Washington, Seattle). Recombinant PLA2s were diluted to solutions of 400 ng/mL and 4 μg/mL concentration, and TR-FIA was performed as described above.
Statistical analysis. We show data as mean ± 95% confidence intervals. Wilcoxon nonparametric signed rank test (SPSS) was used to study the significance of between-group differences in bactericidal activity.
Concentration of PLA2-IIA in acute phase serum. The acute phase serum had high concentrations of PLA2-IIA (278.8 ± 50.0 μg/L; n = 8). From these serum samples, 97.0% of PLA2-IIA was removed by immunoadsorption with rabbit anti-human PLA2-IIA IgG. The residual concentration of PLA2-IIA in the acute phase serum was 8.41 ± 1.25 μg/L. The mouse monoclonal anti-PLA-IIA affinity sorbent removed 99% of PLA2-IIA from serum samples (n = 2). Serum samples of healthy subjects had low concentrations of PLA2-IIA (1.71 ± 0.54 μg/L; n = 6). After immunoadsorption by either depletion method, the PLA2-IIA content of serum of healthy subjects was below the detection limit (∼1 μg/L) of the assay.
Bactericidal activity of PLA2-IIA in serum. To evaluate the antibacterial function of PLA2-IIA in acute phase serum, we measured the bactericidal activity of serum before and after removing PLA2-IIA. Acute phase serum appeared to be highly bactericidal against S. aureus and L. monocytogenes but not against E. coli (figure 1A–1C). Serum of healthy subjects with low concentrations of PLA2-IIA showed antibacterial activity against L. monocytogenes (figure 1D). No bactericidal effect was observed against S. aureus (figure 1E) or E. coli (data not shown) by serum of healthy persons.
The bactericidal effect of acute phase serum against S. aureus was abolished by depleting PLA2-IIA by both mouse monoclonal and rabbit polyclonal anti-recombinant human PLA2-IIA IgG (figures 1A and 2A). The acute phase serum depleted by rabbit polyclonal antibody killed L. monocytogenes but less efficiently than before depletion (figure 1B). The bactericidal effect of acute phase serum against L. monocytogenes was effectively abolished when mouse monoclonal affinity sorbent was used (figure 2B), probably because of a more complete depletion of PLA2-IIA (99% vs. 97%). The bactericidal effect of serum of healthy persons against L. monocytogenes was totally abolished by immunoadsorption by either antibody (figures 1D and 2C). To further confirm this finding, immunoaffinity-purified small intestine PLA2-IIA extract was added to PLA2-IIA-depleted acute phase serum and NHS and tested against S. aureus and L. monocytogenes. Small intestine PLA2-IIA restored the antibacterial effect of acute phase and NHS depleted with PLA2-IIA (figure 2B and 2C). When small intestine PLA2-IIA was added to NHS to equal PLA2-IIA concentration of acute phase serum, the bactericidal activity against S. aureus increased to correspond with that of acute phase serum (figure 2D). Shaking and centrifuging did not change the bactericidal effect of acute phase serum against S. aureus (figure 2E).
Cross-reactivity of polyclonal PLA2-IIA antibody. The recombinant human PLA2-IIA was detected by TR-FIA, and the measured concentration was consistent with the protein concentration determined by an independent UV photometric method (Ultrospec III; Pharmacia LKB Biochrom) by use of 280-nm absorption coefficients calculated, as described elsewhere . The other PLA2 subgroups (PLA2-IB, -IIE, -IIF, -V, -X, and -XII) gave very low signals by TR-FIA (0.01%–1.0%; table 2). The results suggest that the polyclonal antibody used in the current experiments is specific for human PLA2-IIA and that there are negligible cross-reactions with other secretory PLA2s.
Our results indicate that the bactericidal effect of acute phase serum is due to the presence of PLA2-IIA. Bactericidal properties of PLA2-IIA are well known. The mechanism of bacterial killing by PLA2-IIA requires the degradation of bacterial phospholipids by PLA2-IIA . Because of the surface cationic charge, PLA2-IIA can effectively penetrate the highly anionic bacterial cell wall . Weinrauch et al.  reported that human and rabbit PLA2-IIAs kill staphylococci and other gram-positive bacteria in vitro . The LD90 values for purified human PLA2-IIA against S. aureus were 50 μg/L  and 100 μg/L when suspended in baboon serum . Our results correlate fairly well with those results.
Although PLA2-IIA alone can kill gram-positive bacteria [8, 11, 20], the bactericidal mechanism of PLA2-IIA against gram-negative bacteria requires the presence of the bactericidal/permeability-increasing protein  or the components of complement . Unlike our results, Weinrauch et al.  found that baboon acute phase serum killed E. coli with a mechanism independent of PLA2-IIA. This could be explained by species differences between humans and baboons or by the fact that Weinrauch et al. produced acute phase serum by causing sepsis in baboons with intravenous injections of E. coli. Thus, the bactericidal effect of baboon serum against E. coli might have resulted from some other antibacterial agents specific for E. coli.
To study the circulating PLA2-IIA as an antibacterial effector in human bacterial infections, the antibacterial properties of normal and acute phase serum were determined before and after removal of PLA2-IIA. We recently showed that PLA2 groups IIA, IID, IIE, V, X, and XII are bactericidal against gram-positive bacteria, whereas PLA2 groups IB and IIF do not possess marked bactericidal potency [11, 15]. To our knowledge, there are no published data on the presence of other secretory low-molecular- weight PLA2s other than PLA2-IB and -IIA in human serum. Recently, Degousee et al.  showed that the secretory granules of human neutrophils contain PLA2-V and -X. Of these, only PLA2-V is released into extracellular space when neutrophils are exposed to the bacterial tripeptide FMLP or with opsonized zymosan . Considering these findings, it is possible that PLA2-V may be present in human serum during bacteremia, and the bactericidal properties of acute phase serum may partly be explained by the presence of PLA2-V. However, PLA2-V has not been found in acute phase serum . Therefore, it is most probable that PLA2-IIA is responsible for the bacterial killing in acute phase serum.
Our results show that PLA2-IIA, at concentrations prevailing in human circulation in bacterial infections, is able to kill gram-positive bacteria. Eliminating PLA2-IIA from serum abolished the majority of this bactericidal potency. PLA2-IIA may thus play an important role in preventing the spread of bacteria in septic infections. Moreover, PLA2-IIA present at low concentrations in NHS is capable of killing L. monocytogenes.
We previously showed that serum of PLA2-IIA transgenic mice  killed S. aureus , but not E. coli , and that removal of PLA2-IIA abolished the bactericidal activity of transgenic mouse serum against S. aureus . These findings underscore the antibacterial function of serum PLA2-IIA against gram-positive bacteria. The present data support these findings and suggest that PLA2-IIA is an important effector molecule in killing gram-positive bacteria in human circulation in bacterial infections. We also showed earlier that high (nonphysiologic) concentrations of rat PLA2-IIA have a minor bacteriostatic effect against E. coli . It remains unclear how much of the susceptibility of bacteria to PLA2-IIA depends on general structural differences between gram-positive and gram-negative bacteria and how much can be explained by differences between bacterial species and strains. The interactions of PLA2-IIA and other bactericidal proteins in circulation of patients with bacterial infections, such as components of complement and C-reactive protein, remain to be investigated.
In this study, we showed that PLA2-IIA in normal human and acute phase serum effectively kill L. monocytogenes. However, L. monocytes sometimes causes bacteremia, especially in immunocompromised persons. The mechanisms of listerial pathogenesis were thoroughly reviewed recently . L. monocytogenes likely can escape the bactericidal effect of PLA2-IIA by its ability to invade various cells and to rapidly reproduce in the cytoplasm of the host cell. The invasion of endothelial cells by L. monocytogenes is an important event in the pathogenesis of listerial sepsis and other life-threatening infections .
We thank Kari Pulkki for serum samples; Michael H. Gelb for recombinant secretory phospholipase A2; Heikki Peuravuori, Kati Talvinen, and Kirsi Hanikka for technical assistance; Matti Grönroos for help in statistical analysis; and Pentti Huovinen and Jaana Vuopio-Varkila for the bacterial strains.