Studies of the antimicrobial activity of neutrophil defensins have mostly been carried out in microbiological media, and their effects on the host defense in physiological conditions are unclear. We examined 1) the antibacterial activity of defensins in physiological media with and without lung tissue present, 2) the effect of defensins on hydrogen peroxide (H2O2) production by lung tissue that had been exposed to bacteria, and 3) the effect of diphenyleneiodonium (DPI), an inhibitor of reactive oxygen species formation, on the antibacterial activity of defensins in the presence of lung tissue. Defensins were incubated with Escherichia colior Pseudomonas aeruginosa in the absence or presence of primary cultured mouse lung explants. Defensins reduced bacterial counts by ∼65-fold and ∼25-fold, respectively, at 48 h; bacterial counts were further decreased by ∼600-fold and ∼12,000-fold, respectively, in the presence of lung tissue. Defensins induced H2O2 production by lung tissue, and the rate of killing of E. coli by defensins was reduced by ∼2,500-fold in the presence of 10 μM DPI. We conclude that defensins exert a significant antimicrobial effect under physiological conditions and that this effect is enhanced in the presence of lung tissue by a mechanism that involves the production of reactive oxygen species.

  • reactive oxygen species
  • host defense
  • antioxidant
  • NADPH oxidase
  • lung injury

in humans,neutrophil peptide-1, -2, and -3 (HNP-1, -2, and -3, respectively), which are also known as defensins, constitute up to 50% of the total protein within the azurophil granules (7). Defensins are active against gram-positive and gram-negative bacteria in vitro in microbiological media, typically under low-salt conditions (10). The biological effect of defensins in physiological conditions has not been well documented.

Defensins are implicated in a variety of lung inflammatory diseases (1, 13, 19, 24, 25, 28), although their precise role remains poorly defined. Using isolated neutrophils in vitro, it has been reported that cationic proteins including defensins and proteinase 3 of azurophil granules may interfere with the activation of neutrophil NADPH oxidase (30). This effect may severely compromise the phagocytic killing of staphylococci (15). However, it could be argued that isolated neutrophil systems do not reproduce the complex cell-cell and cell-matrix interactions that are seen in the lung. To understand the lung host defense process more completely, models preserving these interactions are preferable.

We have recently developed a rat model of lung explants that consists of the full complement of lung cells (in normal ratios and spatial configuration) that can thus serve as an ex vivo lung model for examining the host response (34). The major advantages of this model are the maintenance of cell-cell and cell-matrix interactions under conditions simulating those seen in the intact lung and the ability to react to exogenous challenges without influence from circulating cells. We adapted this rat lung explant model to mice, because the effects of defensins on murine cells have been shown to be highly comparable to those on human cells (2, 17). The interaction between primary cultured lung parenchymal tissue and purified defensins in this model may more accurately reflect in vivo processes than cell line systems.

In the present study, we tested the antimicrobial effect of defensins alone in physiological media in the presence or absence of cultured lung tissue. We also examined the effect of defensins on hydrogen peroxide production by lung tissue with or without bacteria present. Bacterial killing by defensins was enhanced in the presence of lung tissue, and this effect was attenuated by diphenyleneiodonium (DPI), an inhibitor of NADPH oxidase and thus of reactive oxygen species formation. We demonstrated that defensins exert bactericidal activity not only directly by permeabilizing the bacterial cell membrane but also indirectly by enhancing the oxidative activity of the lung.


Isolation of Defensins

Defensins were isolated from the sputum of patients with cystic fibrosis. The technique for purification of defensins has previously been described (28, 35). In brief, purulent sputum was mixed with 5% acetic acid in a 5-ml aliquot, homogenized in a polypropylene conical tube (Blue Max Jr, Becton Dickinson, Franklin Lakes, NJ), and stirred for 24 h at 4°C. The liquid extracts were cleared by centrifugation at 27,000 g for 30 min. Ten percent dithiothreitol (DTT; BioShop, Burlington, ON) was added to the sputum supernatant for stabilization of defensins during subsequent purification steps. The supernatant was then loaded onto a polyacrylamide gel permeation column of Bio-Gel P-10 (Bio-Rad Laboratories, Hercules, CA) and eluted with 5% acetic acid. Fractions were collected, and defensin-containing fractions were pooled and analyzed for purity by 12.5% acid-urea (AU)-PAGE or 16% tricine-SDS gels as previously described (16) and also by mass spectroscopy (Mass Spectrometry Laboratory, Molecular Medicine Research Center, University of Toronto, Toronto, ON). Fractions corresponding to the defensins were pooled, lyophilized, and resolubilized in 0.01% acetic acid. The protein concentrations were spectrophotometrically measured using absorbance at 280 nm. The theoretical absorbance at 280 nm was calculated from the molar extinction coefficients of tryptophan, tyrosine, and cysteine (10). Working defensin solutions were prepared from a 2 mg/ml stock solution, and a defensin-free vehicle solution was prepared with 0.01% acetic acid.

The lipopolysaccharide (LPS) content levels of the purified defensins were determined (see Assay for LPS). The LPS concentration in the lung explant culture media that contained the highest concentration of 100 μg/ml defensins, which was used for stimulation of lung tissues, was found to be <100 pg/ml.

Preparation of Bacterial Suspensions

An Escherichia coli (E. coli 25922, American Type Culture Collection, Manassas, VA) strain and aPseudomonas aeruginosa (P. aeruginosa 33358, American Type Culture Collection) strain were used. Bacteria were cultured in full-strength trypticase-soy broth (TSB; Becton Dickinson, Cockeysville, MD) overnight (13–14 h) at 37°C in a shaking water bath to obtain stationary-phase organisms as previously described (10). A 2-ml aliquot of the overnight culture was transferred to 50 ml of fresh full-strength TSB and incubated for 2.0 h at 37°C (with shaking) to obtain midlogarithmic phase cells. A portion (25 ml) of the culture was centrifuged for 10 min at 800 g at 4°C. After the supernatant was discarded, the bacterial pellet was resuspended in cold sterile buffer and washed again at 800 g for 10 min at 4°C. The pellet was resuspended in 5 ml of buffer, adjusted to a desired stock concentration in cold sodium phosphate buffer, kept on ice, and mixed (via vortex) before use.

Mouse Lung Explants

Male C57BL/6J mice (Jackson Laboratory, Bar Harbor, ME) of age 8–10 wk and weight 23–25 g were used in accordance with the institutional animal welfare guidelines of the University of Toronto. We adapted the rat lung explant model (34) to mice. In brief, all animals were anesthetized with pentobarbital sodium (40 mg/kg ip) and intubated through a tracheostomy with a sterile 20-gauge angiocatheter (Angiocath, Becton Dickinson Infusion Therapy Systems, Sandy, UT). The surfaces of the anterior chest wall and upper abdomen were sterilized with 70% ethanol. Through a midline abdominal incision, the animals were exsanguinated by dissection of the abdominal aorta. After the chest was opened, the right ventricle was punctured, and a 23-gauge needle (PrecisionGlide, Becton Dickinson) was advanced into the main pulmonary artery. The pulmonary vessels were perfused in situ with 3 ml of cold normal saline while the lung was ventilated via a 1-ml syringe. Using aseptic technique, the trachea, lungs, and heart were dissected en bloc from the animals.

To obtain thin, reproducibly sized pieces of lung tissue, the lungs were prepared for slicing by adapting an agarose inflation technique (34). Low-melting-point agarose (1% wt/vol agarose type VII with low gelling temperature; Sigma Canada, Oakville, ON) was dissolved in bicarbonate-buffered culture medium (containing minimum essential medium, sodium bicarbonate, amino acid supplement, sodium pyruvate, vitamin supplement, gentamicin, and insulin), melted in a microwave oven, and cooled to 37°C before use. The agarose was then instilled as a liquid into the lung through the trachea, which inflated the lung to approximate total lung capacity (2.0–2.2 ml). The lungs were placed in a sterile container at 4°C for 15 min, which solidified the agarose. The lungs were then separated from the heart. Agarose (4%) dissolved in bicarbonate-buffered culture medium melted at 37°C was poured into a sterile, open-mouth, 60-ml B-D syringe (Becton Dickinson, Franklin Lakes, NJ) until the syringe was two-thirds full. The left and right lungs were then separately placed in the syringes and embedded by filling the syringes with 4% agarose. A rubber stopper was used to close the mouth of each syringe. Each syringe containing a lung was cooled at 4°C for 20 min, which solidified the embedding agarose. The resulting agarose-filled and -embedded lungs were then sliced on a microtome (PanaVise Products, Reno, NV) into slices 0.3–0.5 mm thick.

The lung explants were incubated overnight in a 90 × 15-mm petri dish (Nunclon, Nalge Nunc International, Copenhagen, Denmark) with 20 ml of bicarbonate-buffered culture medium at 37°C in a humidified chamber using 5% CO2-95% room air.

The lung explants were then washed extensively with fresh bicarbonate-buffered culture medium and transferred to a preweighed tissue culture multiwell plate (24 flat-bottom, 17 × 16-mm wells, 200 mm2 area per well; Linbro, Flow Laboratories, McLean, VA). Lung slices were cultured in a total volume of 500 μl of antibiotic-free bicarbonate-buffered culture medium that yielded a fluid layer >2.5 mm deep and thus covered the 0.3- to 0.5-mm-thick lung slices. The weights of the lung explants in each well were determined. Experiments were initiated after the culture plate was gently shaken to mix the defensins, bacteria, and DPI in the lung culture medium.

Experimental Protocol

Series I: lactate dehydrogenase release by lung explants.

Cell injury was assessed by measuring the lactate dehydrogenase (LDH) that was released into the culture medium and in homogenates of lung explants in the presence or absence of bacteria at time points of 4, 8, 12, 24, and 48 h. To validate the LDH technique, “positive” control groups were also included by adding 100 pg/ml (the highest possible LPS concentration in the defensin solution) or 100 ng/ml LPS to the culture medium.

Series II: effect of defensins on bacterial growth with no lung explants present.

To examine the effect of defensins on bacterial killing, E. colior P. aeruginosa were added to the bicarbonate-buffered culture medium and incubated with or without the defensins (100 μg/ml) for up to 48 h (n = 6 each from 3 animals) at 37°C. At the end of the incubation period, the culture medium was immediately removed for estimation of viable bacteria.

Series III: effect of defensins on production of hydrogen peroxide and bacterial killing in the presence of lung explants.

Lung explants were randomly divided into six groups, with explants for each group derived from at least three different animals. The groups were treated as follows: group 1: control, no bacteria;group 2: defensins (100 μg/ml); group 3: E. coli; group 4: E. coli + defensins;group 5: P. aeruginosa; and group 6:P. aeruginosa + defensins. The E. coli orP. aeruginosa were added to the bicarbonate-buffered culture medium that contained the lung explants. In each group, the lung explants were incubated for either 10 h or 48 h (n = 6 each from 3 animals) at 37°C. At the end of the incubation period, the supernatants were collected for immediate estimation of viable bacteria, and portions of the supernatants were stored at −70°C for subsequent hydrogen peroxide measurements.

To examine bacterial adherence to the lung explants in the presence of defensins, the lung slices were washed, homogenized in PBS, and incubated in 0.2% Triton X-100 lysis buffer, and the resulting suspensions were plated for estimation of viable bacteria. Lung homogenates were then spun down, and the supernatants were stored at −70°C for later measurement of hydrogen peroxide.

Series IV: effects of DPI on bacterial growth in the presence or absence of lung explants.

In higher organisms, defense against bacterial pathogens is mediated in part by the generation of reactive oxygen species. To examine whether oxidant products are involved in the bacterial killing that is induced by defensins, lung explants were cultured with DPI (10 μM) in the presence or absence of defensins. This concentration of DPI has been shown to inhibit production of hydrogen peroxide, superoxide, and hypochlorous acid by >90% after 10 min of incubation with neutrophils (9) and to reduce ion transport by alveolar epithelial cells when cellular levels of reactive oxygen species were decreased (11). The half-life of DPI has not been well documented but appears to be a matter of a few hours (11). We thus choose a 4-h study period [as described by others using alveolar epithelial cells (11)] to examine the effect of DPI on bacterial killing. It has been shown that the patterns of response to oxidant exposure of E. coli and P. aeruginosa are similar, which suggest common bactericidal mechanisms (5). We thus focused on E. coli in the present study. Lung explants were randomly divided into four groups of six experiments from at least three different animals in each condition and were treated as follows: group 1: E. coli; group 2: E. coli + DPI; group 3: E. coli + defensins; and group 4:E. coli + defensins + DPI.

To examine whether defensins directly interact with DPI to influence bacterial killing, the same experimental groups were also included in the absence of lung explants in culture. In each condition, the culture plates were placed in a shaking incubator at 37°C for 4 h. At the end of the incubation period, the supernatants were collected for immediate estimation of viable bacteria, and portions of the supernatants and lung homogenates were stored at −70°C for subsequent hydrogen peroxide measurement.

Assay for LPS.

The LPS content of the culture medium was determined by the assay of the Limulus amebocyte lysate Pyrochrome (Associates of Cape Cod, Falmouth, MA). Briefly, in the presence of LPS, factors inLimulus amebocyte lysate are activated in a proteolytic cascade that results in the cleavage of a colorless artificial peptide substrate that is present in Pyrochrome lysate. Proteolytic cleavage of the substrate liberates p-nitroaniline, which is yellow and absorbs at 405 nm. The test was performed by adding a volume of Pyrochrome lysate to a volume of defensin solution or culture medium and incubating the reaction mixture at 37°C. The greater the LPS concentration in the sample, the faster p-nitroaniline was produced. Pyrochrome lysate was used to quantify LPS concentration. We used the kinetic method, in which the time taken to reach a particular absorbance at 405 nm (the onset time) was determined; higher LPS concentrations gave shorter onset times. Standard curves were constructed by plotting log onset time against log concentration of the LPS standards, and these curves were used to calculate LPS concentrations in the defensin solutions and culture medium.

The LPS levels detected from lung culture medium 10 h after bacterial challenge were similar in the absence (3.7 ± 0.4 ng/ml in E. coli group; 6.7 ± 0.2 ng/ml in P. aeruginosa group) or presence (3.0 ± 0.3 ng/ml in E. coli group; 6.9 ± 0.1 ng/ml in P. aeruginosagroup) of defensins, which suggests that any observation that resulted from this model was not directly related to LPS release.

Assay for LDH.

Lung explant viability is expressed as the percent LDH content present in the supernatant compared with the total LDH, and total LDH was determined as the sum of the supernatant LDH and the tissue LDH derived from the explant homogenate. Briefly, lung explants were washed three times in PBS and homogenized in 2 ml of PBS. An equal volume of 0.2% Triton X-100 lysis buffer was added to the homogenate suspensions, which were then incubated at room temperature for at least 10 min. A colorimetric assay was performed using an LDH cytotoxicity detection kit (Boehringer Mannheim) as we have previously described (34).

Assay for hydrogen peroxide.

Hydrogen peroxide concentrations from the supernatant and the lung homogenates were measured using a colorimetric assay (Bioxytech H2O2-560, Oxis International, Portland, OR). The assay is based on the oxidation of ferrous ions (Fe2+) to ferric ions (Fe3+) by hydrogen peroxide under acidic conditions (Fe2+ + H2O2 → Fe3+ + ·OH + OH). The Fe3+ binds with the indicator dye xylenol orange (XO, a sodium salt) to form a stable, colored complex that can be measured at 560 nm (Fe3+ + XO → Fe3+—XO). A standard curve was built according to the instructions provided with the kit, and concentration values were normalized per gram of tissue weight.

Statistical Analysis

A two-way ANOVA and subsequent Tukey-Kramer test were used for statistical analysis of the data. Differences were considered statistically significant at P < 0.05. Data are presented as means ± SE.


Purification and characterization of neutrophil defensins.

The AU-PAGE resulted in a major band of purified defensins (isolated from an acid sputum extract by gel filtration on a Bio-Gel P-10 chromatograph column) that was similar in electrophoretic mobility and appearance to that of a mixture of commercially synthesized human defensins, namely, HNP-1 and HNP-2 (Sigma Chemical, St. Louis, MO) (Fig. 1 A). Mass spectroscopy also demonstrated that the masses of the purified defensins were in accordance with those of HNP-1, -2, and -3 (Fig.1 B), which were previously documented (10).

Fig. 1.

A: analysis via acid-urea-PAGE of defensins isolated from sputum of patients with cystic fibrosis. Defensins [human neutrophil peptides (HNP-1, -2, and -3); lane b] were isolated from an acid sputum extract (lane a) by gel filtration on a Bio-Gel P-10 chromatograph column. Lane c: commercially synthesized HNP-1 and HNP-2. B: analysis by mass spectroscopy of purified defensins from sputum of patients with cystic fibrosis. Nos. above peaks, masses [mass-to-charge ratio (m/z)] of purified defensins. Documented masses for HNP-1, -2, and -3 are 3,442, 3,371, and 3,486, respectively.

Validation of murine model of lung explants.

To investigate the bacterial killing function of defensins in the lung, a model of primary cultured mouse lung explants was developed. The LDH assays demonstrated that the control lung explants and those exposed to very low concentrations of LPS (100 pg/ml) or to concentrations of inoculated bacteria remained viable over the study period. There was no cell injury in these explants as assessed by LDH release. Supernatant LDH represented ∼3% of the total LDH (Fig.2 A), which is comparable to the values reported for isolated viable lung epithelial cells (22). In contrast, in explants exposed to a high concentration of LPS (100 ng/ml; Fig. 2 B), the supernatant LDH was much higher, which suggests an active cellular response to LPS stimulation by the primary cultured lung explants. This observation further confirms that the percent LDH that is in the supernatant compared with the total LDH concentration is a reliable index of cell injury as has been previously reported (20, 22). The LDH levels tended to be higher in the bacteria-treated lung explants compared with the control group, but the difference was not statistically significant (Fig. 2 A).

Fig. 2.

A: cytotoxicity of E. coli on lung explants is expressed as the percent lactate dehydrogenase (LDH) content in the supernatant of the total LDH over time. Total LDH was determined as the sum of supernatant LDH and LDH from the homogenates of explants. LDH release tended to be higher in the bacteria-treated lung explants compared with the control group, but the difference was not statistically significant. B: at 24 h, LDH release markedly increased in the explants treated with the higher concentration of lipopolysaccharide (LPS; 100 ng/ml) but not with the lower concentration (100 pg/ml). *P < 0.05 vs. 100 ng/ml LPS.

Antimicrobial effect of defensins in the physiological culture medium.

It is accepted that defensins are naturally expressed antimicrobial molecules. However, most studies have been performed in microbiological media that may not be suitable for lung tissue culture or cell growth. To determine whether this antimicrobial activity occurs under physiological conditions of varying pH and salt concentrations, we examined the effect of defensins (100 μg/ml) in a standard tissue culture medium. After a 48-h incubation of the same concentration of bacteria, the E. coli counts increased by 10-fold, andP. aeruginosa increased by 30-fold in the control groups (Fig.3 A). Defensins showed significant antibacterial effects: the number of E. colidecreased by ∼65-fold, and P. aeruginosa decreased by ∼25-fold compared with the vehicle control groups (Fig.3 A). These data confirm that defensins possess bactericidal activity in physiological media and suggest greater activity towardE. coli than P. aeruginosa.

Fig. 3.

Antibacterial activity of defensins (100 μg/ml) on bacterial killing activity was examined in standard tissue culture medium after 48-h incubation with bacteria [3 × 105colony-forming units (cfu)/ml]. A: antibacterial activity in the absence of lung explants (−). B: activity in the presence of lung explants (+). Antibacterial effect of defensins was more pronounced in the presence of cultured lung tissue than in its absence. Open bars, bacteria inoculated; hatched bars, bacteria recovered at 48 h; solid bars, bacteria + defensins at 48 h. *P < 0.05 vs. bacteria + defensins in the same culture conditions. †P < 0.05 vs. other groups in the same culture conditions. C: number of colony-forming units of E. coli recovered from lung homogenates after extensive lung tissue washing in PBS. Number of bacteria that adhered to the lung tissue was approximately fivefold lower in the defensin-treated group than in the control group (P = not significant).

Lung tissue enhances the antimicrobial effect of defensins.

Lung explants were incubated with the same amount of bacteria under the physiological culture conditions described. Defensins decreased theE. coli counts by ∼600-fold and the P. aeruginosa counts by ∼12,000-fold at 48 h (Fig.3 B). It is evident that the antibacterial effect of defensins was more pronounced in the presence of cultured lung tissue than in its absence (Fig. 3, A and B).

To examine whether the decreased bacterial count seen in the supernatant resulted from an increased adherence of viable bacteria to the lung explants by defensins, the lung tissue was homogenized and immediately plated for E. coli counting. Figure3 C shows that the bacterial count from the lung homogenates was approximately fivefold lower in the defensin-treated group than in the control group, although the difference did not reach statistical significance.

Oxidant products from lung explants contribute to enhanced antibacterial effect of defensins.

The enhanced antibacterial effect of defensins in the presence of lung explants suggests that defensins may induce additional bactericidal factors from lung cells. We speculated that one component of the host defense activity of defensins is to regulate production of oxidant products in lung tissue. We first investigated whether reactive oxygen species are upregulated in lung tissue after defensin (100 μg/ml) stimulation. Figure 4 shows that hydrogen peroxide concentration was significantly increased 4 h after defensin administration compared with that in the vehicle control group. Higher concentrations (>100 μg/ml) of defensins were not tested in the present study, because they can have effects other than bacterial killing, such as direct cytotoxicity and the induction of cytokine release (24, 32).

Fig. 4.

Time-dependent production of hydrogen peroxide (H2O2) by lung explants that were incubated with 100 μg/ml defensins in the absence of bacteria at 37°C. *P < 0.05 compared with time 0 within group. †P < 0.05 compared with vehicle (defensin-free buffer) at the same time points.

DPI is an inhibitor of NADPH oxidase, which attenuates reactive oxygen species formation. To determine whether the increased lung oxidant response induced by defensins was related to the enhanced bacterial killing activity, lung explants were separately challenged with E. coli in the absence or presence of defensins or DPI. DPI did not affect the number of colony-forming units when incubated with E. coli alone, which indicates that DPI did not have a direct antibacterial effect (Fig.5 A). Defensins dramatically decreased the bacterial number associated with an increased hydrogen peroxide concentration (Fig. 5, A and B). The rate of bacterial killing by defensins was significantly blunted when DPI was administered (Fig. 5 A). This effect was not due to a direct interaction between defensins and DPI, because no additive bacterial killing was seen when defensins and DPI were incubated together with E. coli in the absence of lung explants (Fig.6).

Fig. 5.

A: number of colony-forming units of E. colirecovered from lung culture supernatant. B: concentration of total H2O2 by lung explants. Total H2O2 was determined as the sum of H2O2 concentrations from the supernatant and the explant homogenates. Lung explants were incubated for 4 h withE. coli alone (open bars), E. coli + 10 μM diphenyleneiodonium (DPI; hatched bars), E. coli + 100 μg/ml defensins (crosshatched bars), or E. coli + 100 μg/ml defensins + 10 μM DPI (solid bars). *P< 0.05 compared with E. coli or E. coli + DPI. †P < 0.05 compared with E. coli+ defensins. ‡P < 0.05 compared with E. coli + DPI or E. coli + defensins + DPI.

Fig. 6.

Number of colony-forming units of E. colirecovered from supernatant without lung explants in culture. To examine whether defensins directly interact with DPI to influence bacterial killing, E. coli were incubated for 4 h either alone (open bar) or with 10 μM DPI (hatched bar), 100 μg/ml defensins (crosshatched bar), or 100 μg/ml defensins + 10 μM DPI (solid bar). *P < 0.05 compared with E. coli orE. coli + DPI. There was no direct interaction between defensins and DPI on bacterial killing.

The LDH content of the supernatants was similar, ranging from 5.1 to 5.5% of the total LDH in the groups of E. coli alone,E. coli + DPI, and E. coli + defensins + DPI. This suggests that DPI at the concentration used did not have a cytotoxic effect on the lung explants. The lung tissue that was treated with E. coli + defensins showed a decreased LDH release (3.37 ± 0.05% of total LDH), although the difference did not reach statistical significance.


To date, most defensin studies to test antibacterial effects have been performed in media conducive to microbial growth (typically low-salt, low ionic strength solutions). The present study demonstrates for the first time that defensins exert antimicrobial activity in standard tissue culture conditions of physiological pH, temperature, and salt concentration. The bacterial counts decreased in the defensin-treated group compared with the vehicle group at 10 h (data not shown) and decreased further at 48 h, which suggests that the antimicrobial activity of defensins is long lasting under these conditions.

We also investigated the effect of defensins and lung tissue together on bacterial killing in physiological media. In the presence of cultured lung tissue, the maximal killing capacity of defensins was >1,000-fold greater than in its absence, which indicates that the antibacterial activity is further enhanced by lung tissue. Theoretically, the lower bacterial number seen in the culture supernatant might be due to an increased bacterial adherence to lung tissue in the presence of defensins. We excluded this possibility by showing that the number of bacteria recovered from the lung homogenates was actually slightly lower in the defensin-treated group than in theE. coli-alone group. Our data thus demonstrate that under physiological conditions, cationic defensins kill bacteria not only directly via the well-described mechanisms of increasing membrane permeability and cell lysis (7) but also indirectly by producing bactericidal products from lung tissue.

Although phagocytes are the main source of antibacterial reactive oxygen species within the host, other sources of oxidants may also contribute to bactericidal activity. In particular, lung endothelial and epithelial cells produce oxidant products in response to a number of stimuli including inflammation. We and others have previously demonstrated that stimulated lung cells or tissue produce reactive oxygen species such as hydrogen peroxide through the assembly of the reduced NADPH oxidase complex (8, 26, 33, 34). In the present study, we report for the first time that defensins directly induce production of hydrogen peroxide by lung explants. The generation of oxidant products by lung tissue after defensin stimulation may play an important role in host defense. It has been shown that hydrogen peroxide dramatically reduces the growth rate of E. coli in culture without cytotoxicity to cultured fibroblasts (12). There are a number of ways in which hydrogen peroxide may inhibit bacterial growth. For example, bacteria may react to oxidative stress by invoking two distinct oxidant responses: the peroxide stimulon and the superoxide stimulon. The two stimulons each contain genes that constitute the OxyR or SoxRS regulon, respectively (6). Activation of these genes inhibits cell division (12). In addition, hydrogen peroxide induces DNA damage in E. coli; this process is mediated by a Fenton reaction that generates hydroxyl radicals from hydrogen peroxide (14).

To determine the role of oxidant mechanisms in the killing of E. coli by defensins, the effect of DPI on the production of hydrogen peroxide and the bacterial count was measured. DPI inhibits the action of NADPH oxidase on the flavoprotein in blocking the sequence of NADPH → FAD protein → cytochrome b → reactive oxygen species (4). At a concentration of 10 μM, DPI has been shown to inhibit superoxide, hydrogen peroxide, and hypochlorous acid production (9). We measured only hydrogen peroxide in the present study for two reasons. First, it appears that hydrogen peroxide is a sensitive indicator of oxidant production. Hampton and Winterbourn (9) observed that certain stimuli lead to production of hydrogen peroxide without measurable superoxide and that this process is more sensitive to DPI. Second, the inhibition of hydrogen peroxide production by DPI is as sensitive as that for hypochlorous acid production (9). Hydrogen peroxide is thus considered to be a sensitive marker of oxidant production. Our data show that the rate of E. coli killing by defensins was reduced 2,500-fold in the presence of DPI. Our study clearly demonstrates that the generation of oxidants induced by defensins contributes to enhanced bacterial killing in the lung.

The increased oxidant production and NADPH oxidation by defensins seen in the present study is somewhat contradictory to the data obtained using cell-free systems or cultures of isolated neutrophils in which NADPH oxidase and superoxide production were reportedly decreased (15, 29). There are a number of possible explanations for the discrepancies between these other studies and our own. First, the time courses are different: our study lasted up to 4 h as opposed to only 5 min in the study by Tal and Aviram (29). We observed that defensins induced hydrogen peroxide production in a time-dependent fashion. Second, we used a higher concentration of defensins (100 μg/ml) than were employed in the earlier studies (5–35 μg/ml) (15, 29). Third, the culture medium compositions were different. We used a plasma-free medium, whereas plasma was added in the media used by others (15). It is suggested that proteinases released into plasma may interfere with some activities of defensins (30). Fourth, the most striking and important difference is the models used. We employed a lung explant model in which most blood-derived cells were removed by flushing the pulmonary vasculature, whereas other investigators used isolated neutrophils (15, 29, 30). It is known that the neutrophil azurophil granules can release several substances that are inhibitory to NADPH oxidase. Tal and colleagues (30) have recently demonstrated that proteinase 3, a serine protease distinct from defensins, inhibits neutrophil NADPH oxidase activation and results in decreased superoxide production. It thus appears that between lung parenchymal cells and phagocytes, the immune responses to defensin stimulation may be quite incomparable.

Mechanisms other than oxidant production may be involved in the enhanced bacterial killing activity of defensins in the presence of lung tissue. Defensins may increase secretion of bactericidal compounds (including epithelium-derived proteins and airway surface liquid) that might have significant antibacterial properties and effects on the pulmonary immune response (3, 21, 23, 31).

In summary, purified defensins have significant antimicrobial activity in physiological conditions. This antimicrobial activity is further enhanced by lung tissue through the release of bacteria-killing products including reactive oxygen species. Thus defensins may have an important role in modulating lung host defense through NADPH oxidation in addition to direct antibacterial actions.


We thank Dr. Brian Kavanagh (Department of Critical Care Medicine, Hospital for Sick Children, University of Toronto, Toronto, Ontario, Canada) for performance of the lipopolysaccharide assays.


  • This work was supported by Canadian Institutes of Health Research (CIHR) Grants MOP-44093 (to H. Zhang), MT-13270 (to M. Liu), and MA-8558 (to A. S. Slutsky) and Ontario Thoracic Society Grant-In-Aid 2000-2001 (to H. Zhang and A. S. Slutsky).

  • H. Zhang is a CIHR Fellow. M. Liu is a CIHR New Investigator.

  • Address for reprint requests and other correspondence: H. Zhang, Dept. of Anaesthesia, Univ. of Toronto, Medical Sciences Bldg., Rm. 7334, 1 King's College Circle, Toronto, ON M5S 1A8, Canada (E-mail: haibo.zhang{at}utoronto.ca).

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.


View Abstract