Cytotoxicity towards human endothelial cells, induced by neutrophil myeloperoxidase: protection by ceftazidime

We investigated the effects of the antibiotic ceftazidime (CAZ) on the cytolytic action of the neutrophil myeloperoxidase–hydrogen peroxide–chloride anion system (MPO/H2O2/Cl−). In this system, myeloperoxidase catalyses the conversion of H2O2 and CI− to the cytotoxic agent HOCl. Stimulated neutrophils can release MPO into the extracellular environment and then may cause tissue injury through direct endothelial cells lysis. We showed that human umbilical vein endothelial cells (HUVEC) were capable of taking up active MPO. In presence of H2O2 (10−4 M), this uptake was accompanied by cell lysis. The cytolysis was estimated by the release of 51Cr from HUVEC and expressed as an index of cytotoxicity (IC). Dose dependent protection was obtained for CAZ concentrations ranging from 10−5 to 10−3 M;this can be attributed to inactivation of HOCl by the drug. This protection is comparable to that obtained with methionine and histidine, both of which are known to neutralize HOCl. This protection by CAZ could also be attributed to inactivation of H2O2, but when cytolysis was achieved with H2O2 or O2- generating enzymatic systems, no protection by CAZ was observed. Moreover, the peroxidation activity of MPO (action on H2O2) was not affected by CAZ, while CAZ prevented the chlorination activity of MPO (chlorination of monochlorodimedon). So, we concluded that CAZ acts via HOCl inactivation. These antioxidant properties of CAZ may be clinically useful in pathological situations where excessive activation of neutrophils occurs, such as in sepsis.


Introduction
Sepsis, septic shock, severe trauma, hypovolaemic shock and acute pancreatitis are examples of severe pathological situations accompanied by an intense inflammatory reaction involving activation of specific cells and release of mediators. This inflammatory reaction occasionally overwhelms the organism's defences, with excessive activation of polymorphonuclear leucocytes (PMN). This can lead to the systemic inflammatory response syndrome (SIRS) and eventually to multiple organ dysfunction syndrome (MODS). -PMN activation produces activated oxygen species (mainly superoxide anion and hydrogen peroxide, H202) inflammatory mediators, and proteolytic and hydrolytic enzymes. 4'5 PMN also release myeloperoxidase (MPO) from their granules; in the presence of H202 and chloride anion (Cl-), MPO generates hypochlorous acid (HOC1). HOC1 reacts with amines to form chloramines, but also with H202 itself to produce singlet oxygen, an activated form of oxygen. <7 These activated oxygen species and oxidant products of MPO activity normally play a beneficial role in host defence against invading microorganisms. However, their excessive production can be detrimental for tissues, especially for endothelial cells, such as, when PMN are trapped in capillaries and become strongly adherent to endothelium. 8 They are also destructive for plasma proteins, especially those with thiol (-SH) functions, such as (x 2 macroglobulin, an essential plasmatic antiproteinase. 9-In these situations associated with excessive activation of PMN, therapeutic agents capable of neutralizing these active oxygen species and protecting endothelial cells from oxidant stress would have potential clinical utility. Recent  Assay for MPO enzyme activRy: MPO was purified from human polymorphonuclear neutrophils as described previously. 9'2 Its enzyme activity was determined by spectrophotometric methods. MPO peroxidative activity. MPO peroxidative activity was assayed by measuring the absorbance increase at 460nm caused by the oxidation of ortho-dianisidine. 2 MPO (dissolved in 50 I.tl) was added to 3 ml of 50 mM ortho-dianisidine in S6rensen buffer pH 5.5 and the reaction was started by the addition of H202 at a final concentration of 0.15 mM. The absorbance increase at 460nm was followed for 1 min. One unit of activity was defined as the amount of MPO which produced an absorbance increase of 1 optical density unit per minute.
MPO-dependent chlorination activity. MPOdependent chlorination activity was measured by following the conversion of monochlorodimedon to dichlorodimedon at 290nm. 22 Typical experimental conditions were as follows: MPO (5/.tg) was incubated in 3ml of monochlorodimedon (24btM) in 100mM phosphate buffer (pH 5.5) with NaCl (50mM) added. The reaction was started by the addition of H202 (0.1 mM) and the decrease of absorbance at 290 nm was followed for 1 min. One unit of activity was defined as the amount of MPO which produced an absorbance decrease of 1 optical density unit per minute.
To determine the effects of CAZ on MPO enzyme activities, MPO was incubated with different concentrations of CAZ for 5 min at room temperature before the enzyme assays. The data represent arithmetic means _ _ _ S.D. of triplicates.
MPO uptake by HUVEC. Confluent cells in 1 ml of Hanks' Balanced Salt Solution (HBSS)were incubated at 37C with increasing amounts of MPO (from 0 to 30 l.tg) for 3 h or with a fixed amount of MPO (5t.tg) for periods of time ranging from 0 to 360 min. Each assay was done in triplicate and the entire experiment was repeated three times. After these incubations, cells were washed three times with HBSS, and the MPO content of the cells was estimated by measurement of MPO peroxidative activity (see above). Ortho-dianisidine (lml) was added to adherent cells and the reaction was started by the addition of H202 at a final concentration of 0.15 mM. After an incubation of 30 min, the reaction buffer was collected and the absorbance at 460 nm was read. The amount of MPO present in the cells was calculated from a standard curve for which known amounts of MPO were incubated under the same conditions. Cytotoxiiy assay: Cytotoxicity was assessed by measuring the release of previously incorporated 51Cr. 23 Confluent HUVEC in six multiwell plates were labelled overnight by the addition of 20 btCi 5Cr in culture medium per well (sodium chromate, Amersham). HUVEC were washed in HBSS to remove unincorporated 51Cr and the cells were then submitted to oxidant stress as described below. Each assay was done in triplicate. At the end of the oxidant stress period, the supernatants were collected and the cells were washed three times with HBSS. Supernatant and washings were pooled and 51Cr release JR] was quantified by 7 counting.
Cells were lysed in NaOH (1 N) and the intracellular 5Cr [ 15 min of incubation 0.177 +__ 0.037 lag of MPO were already present in the cells. The maximal Cytotoxicity induced by the xanthine/xanthine uptake reached a plateau after 180 min, with oxidase system. HUVEC in l ml of HBSS were 0.553_ 0.0121.tg taken up. This maximal value treated with 2mM xanthine/20mU of xanthine corresponds to an uptake of 11.00 _-+ 0.24% of oxidase (Boehringer-Manheim) for 2 h following added MPO.
The dose-response curve of MPO uptake by HUVEC is shown in Fig. 2 24 25 and O2 scavengers. 5Cr labelled HUVEC in l ml of HBSS supplemented with one of the Cytotoxiciy of the MPO/HeOe/C1system on putative protectors (CAZ, methionine, histidine HUVEC.. When HUVEC were incubated for 2h or dabco) were incubated for 2h prior to initi-with MPO (51.tg) and H202 (10-4M) added ating the cytotoxic treatments described above, simultaneously, only a low cytotoxicity For the oxidant stress with MPO/H202/CI-, the (IC 5.04 _ _ . 2.1)was found. No cytotoxicity was protector was added with MPO 2h before the obtained with either MPO or H202 alone. To addition of H202 After the stress, IC was deterpermit the uptake of MPO prior to initiating the mined, and the % protection was obtained oxidant stress with H202, the cells were preaccording to the following formula: incubated with MPO for 2h. This time of pre-  Protection of HUVEC from MPO/H202/CIor NaOCI oxidant stresses: To test the effects of putative protectors (CAZ, methionine, histidine, or dabco), the protector was added at different concentrations on HUVEC, together with 5 t.tg of MPO, and the cells were incubated for 2 h prior to addition of H202 (Fig. 4). We controlled the effects of CAZ on MPO uptake by endothelial ceils and found neither inhibiting nor enhancing effects of the antibiotic: without CAZ, we found an uptake of 0.476 _+ 0.0441Ltg MPO, and with CAZ (10-M), an uptake of 0.469 +__ 0.011btg (no statistical difference). After the stress, IC was calculated and the results were expressed as percent protection according to the formula defined in Materials and Methods: a protection of 100% corresponds to a total inhibition of stressinduced cytotoxicity.
The protection by CAZ was dose-dependent and articularly effective at the concentration of 10-M (percent protection 92.4 _ _ _ 1.7). This protection was confirmed by light microscope observations, which indicated a similar cellular aspect for control cells (without stress) and cells  ference between the two compounds), but histi-FIG. 5. Effects of CAZ on MPO enzyme activities. MPO was preincubated with different concentrations of CAZ for 5 min at dine was less effective (p < 0.05 for each room temperature. I-I, peroxidative activity of MPO; N, chlorinaconcentration compared to CAZ and methiotion activity of MPO. MPO enzyme activities obtained in the nine). In contrast, no protection was obtained absence of CAZ were taken as the reference (100%). Each point represents the mean 4-S.D. of two experiments performed in with dabco (percent protection 6.6 -t-9.0). triplicate. *No statistical difference (p > 0.05)and **statistical dif-From these results, we concluded that CAZ ference (p < 0.05) vs. control when analysed with the two-tailed was capable of neutralizing cytotoxicity induced Student's t-test. by the MPO/H202/CIsystem by inactivating the HOC1 generated by MPO. To confirm the direct effect of CAZ on NaOC1, 5aCr labelled HUVEC Effects of CAZ on MPO enzyme activities.. Even at were exposed to NaOC1 stress in presence of concentrations of CAZ (10-3M) which were CAZ, methionine, or histidine; the cells were highly protective for HUVEC submitted to MPO/ incubated at 37C for 2h in the presence of H202/CIstress, only weak inhibition of MPO putative protector (at a concentration of peroxidative activity was observed (90.0 -t-0.7% 10-3 M). NaOC1 (10-3 M) was then added and of control activity) (Fig. 5). In contrast, MPOthe IC was determined after a further 2 h incuba-dependent chlorination of monochlorimedon tion. The IC in the absence of protector was was very sensitive to CAZ: no chlorination was 14.86-t-3.15 and the percent protection, calcuobserved in the presence of 5 x 10-4M and lated as previously described, are presented in only 61.0 -t-4.9% of activity was conserved in the Table 1. Marked protection was seen with CAZ presence of 2.5 x 10-4M CAZ. and methionine, while histidine was weakly active.

Absence of a protective effect of CAZ on oxidant
In this study we demonstrate that endothelial stress generating He02: To test the effect of CAZ cells are capable of binding and taking up myeon HUVEC cytotoxicity induced by RiO2 or O2-, loperoxidase. This uptake shows steep time we incubated the cells in the presence of enzydependency at 37C, and is maximal at 3 h incumatic systems generating this reactive oxygen bation. The yield of the uptake process is low, species: glucose/glucose oxidase or xanthine/ with a maximum incorporation of 11% when xanthine oxidase systems. 5 lag of MPO are added to the culture medium For the two stresses, no statistically significant (Fig. 1). Increasing the amount of MPO added to difference was observed (two-tailed Student's the medium does not increase uptake, which test) between the cytotoxicity obtained in the peaks at 7.5lag (62.4nM) added MPO. At this absence or in the presence of CAZ (10-M). level, 7.8% of the added dose is incorporated The IC values obtained for glucose/glucose after 3h incubation. These results agree with oxidase stress were respectively 25.02 _ _ _ 3.91 those of Zabucchi et al. 26  workers also showed that the enzyme was present on the cell membrane and in the cytoplasm. Using the same cytochemical reaction, we showed uptake of MPO by endothelial cells, and its presence in intracellular granules (results not shown). Spectrophotometric measurement of intracellular enzymatic activity confirmed uptake and showed that the internalized enzyme retained its activity. We used lower concentrations of MPO than Zabucchi et al., but in our study, we used a 3 h incubation (as opposed to the 10 min used by Zabucchi's group). This duration of incubation did not alter the enzyme, because at the end of the period, the sum of enzyme left in the supernatant and that incorporated into the cells exactly equalled the total initially added.
In the presence of H202, MPO produces a cytolytic oxidative stress on endothelial cells, which we have expressed as a cytotoxicity index. This IC increases linearly with the dose of MPO with a plateau at 5 btg per 4 x 105 cells. No cytotoxicity is observed, however, if MPO is not preincubated with the cells prior to addition of H202 This indicates that incorporation of the enzyme (or its absorption to the cell surface) is necessary. In subsequent experiments of the protective effects of added substances, we used a concentration of 5 g for 4 x 105 cells. We chose a 2 h pre-incubation at 37C, because this timing allowed reproducible uptake across batches of cells. After addition of H202, a further 2 h incubation was chosen because this timing allowed a convenient and reproducible cytotoxicity.
Several antibiotics have been reported to have antioxidant properties against active forms of oxygen. 14'27-29 We chose to use ceftazidime because it is a highly effective agent, a broadspectrum ]3-1actamase resistant antibiotic with anti-Pseudomonas activity, widely used in patients in the intensive care unit. CAZ protects endothelial cells from the oxidant stress induced by MPO in a dose dependent fashion, comparably to methionine, and more powerfully than histidine. These two amino acids are capable of scavenging HOCl, formed by the enzyme action of MPO in the presence of H202 and el-. 4'5 The protective effect of CAZ could thus be attributed to its reaction with HOCl, neutralizing the oxidant activity responsible for the cytolysis seen in our experimental model. This mode of action is confirmed by the observation of a protective effect of CAZ on endothelial cells against direct attack by added NaOC1. This protection is less than that of methionine, but superior to that of histidine. This latter compound is less active on NaOC1 stress than on MPO/H202/Clstress. This can be explained by the differences in the two types of stress. In the first case, the concentration of 442 Mediators of Inflammation Vol 4 1995 HOCl directly reaches 10-M. The activity of MPO induces a progressive release of low concentrations of HOC1, without reaching a final concentration of 10-3M. The reaction of CAZ with HOCl occurs simultaneously with its formation by MPO, and prevents its chlorination of monochlorodimedon. On the other hand, CAZ has no effect on the peroxidase activity of MPO. The reaction of CAZ with HOCl could occur at either thioether groups or near amine groups of the antibiotic. It should be noted, however, that according to Lapenna et al., 4 the reaction of CAZ with HOCl does not lead to formation of chloramines, which pleads against involvement of amine groups. The protective effect of methionine confirms a role of sulfur atoms, while histidine's action can only be explained by the presence of nitrogen, with the formation of nontoxic chloramines. The lack of a sulfur atom in histidine could partially explain the lower protective effect of this compound compared to CAZ and methionine.
In our model, neutralization of HOCl by CAZ is the principal mode of action of the antibiotic.
The simultaneous presence of H202 and HOC1 during the activity of MPO could lead to the formation of singlet oxygen (O2), a highly reactive 30 32 species.
We have previously shown that CAZ is capable of interacting with O2, deactivating it more powerfully_ than dabco, a well-characterized quencher. 5-7'25 Under our experimental conditions, dabco at a concentration of 10-M was only weakly protective. This would seem to indicate that singlet oxygen is not the principal source of oxidant stress on endothelial cells induced by the activity of MPO. On the other hand, the lack of efficacy of CAZ against oxidant stress induced by glucose oxidase, and especially xanthine oxidase, which both produce superoxide anion and H202, indicates that the antibiotic does not act directly on these two activated species of oxygen, what has been already demonstrated for other antibiotics. Cantin et al. described similar cytotoxicity of the MPO/H202/Clsystem against epithelial cells, using 2.5 l.tg/ml MPO, and observed a protective effect of CAZ at concentrations approximately equal to those we used. Ottonello et al. 2 showed protection by several antibiotics, including CAZ, against the action of stimulated PMN on lymphoblastoid cells (Daudi cell line), and attributed this effect to inactivation of HOCl produced extracellularly by MPO. Preliminary results from our laboratory suggest that CAZ also protects endothelial cells from the cytolytic activity of activated PMN.
We thus see that CAZ, in addition to its antibiotic activity possesses antioxidant properties.
These could be useful in situations where excessive PMN activation, production of activated species of oxygen, and liberation of MPO are seen. Sepsis and septic shock are two such situations. In our model, the lowest concentration associated with a protective effect (10-5M 5.461.tg/ml, 26% protection) is inferior to the serum concentrations attained in patients treated with this antibiotic: following an intravenous infusion of 2g CAZ in healthy volunteers and patients, the serum concentration peaks at 59 to 83btg/ml. 34 The combination of an antioxidant activity with those of an antibiotic in one molecule allows simultaneous treatment with a well characterized, widely used drug, of both the infectious aspect of the disease process, as well as the consequences of excessive PMN activation.