Lactoferrin decreases LPS-induced mitochondrial dysfunction in cultured cells and in animal endotoxemia model

Abstract

Lactoferrin is a non-heme iron-binding glycoprotein, produced by mucosal epithelial cells and granulocytes in most mammalian species. It is involved in regulation of immune responses, possesses anti-oxidant, anti-carcinogenic, anti-inflammatory properties, and provides protection against various microbial infections. In addition, lactoferrin has been implicated in protection against the development of insult-induced systemic inflammatory response syndrome (SIRS) and its progression into septic conditions in vivo. Here we show a potential mechanism by which lactoferrin lessens oxidative insult at the cellular and tissue levels after lipopolysaccharide (LPS) exposure. Lactoferrin pretreatment of cells decreased LPS-mediated oxidative insults in a dose-dependent manner. Lipopolysaccharide-induced oxidative burst was found to be of mitochondrial origin, and release of reactive oxygen species (ROS) was localized to the respiratory complex III. Importantly, lactoferrin nearly abolished LPS-induced increases in mitochondrial ROS generation and the accumulation of oxidative damage in the DNA. In vivo, pretreatment of experimental animals with lactoferrin significantly (P<0.05) lowered LPS-induced mitochondrial dysfunction as shown by both decreased release of H₂O₂ and DNA damage in the mitochondria. In contrast, deferoxamine, an iron chelating compound, provided only partial protection in LPS-treated animals. Together, these data suggest that lactoferrin protects against oxidative insult at the mitochondrial level, and indicate a potential utility of lactoferrin in prevention and treatment of SIRS.

Introduction

Lactoferrin is an 80-kDa, non-heme iron-binding glycoprotein that belongs to the transferrin family.¹ It is found in most mucosal sites and secondary granules of neutrophils in mammals.^2–4 Several functions have been attributed to lactoferrin as a key component in the host’s first line of defense, contributing to a variety of physiological changes at both the cellular and organ levels.^4,5 Lactoferrin is a major pleiotropic mediator that plays an important role in the development of inflammatory responses.^6–8 It was recently demonstrated that lactoferrin has the ability to inhibit progression of systemic inflammatory response syndrome (SIRS) into sepsis in endotoxemic mice.^9,10 Lactoferrin is well documented as having direct antimicrobial activity, including an iron-dependent bacteriostatic property and a non-iron-dependent bactericidal action on LPS-bearing Gram-negative bacteria.^11,12 While suppressing microbial growth, lactoferrin also exerts its first-line of defense activity with significant impact on the development of adaptive immune responses.¹³ Sequestration of iron by lactoferrin not only inhibits microbial growth but also reduces oxidative stress. In addition, lactoferrin induces anti-allergic responses in animals challenged with ragweed pollen or dust mite proteins as shown by decreased accumulation of inflammatory cells in airways and subepithelium of sensitized experimental animals.¹⁴ Results also show that lactoferrin decreased pollen-induced cellular ROS levels in bronchial epithelial cells and prevented both development of mucin-producing cells as well as levels of mucin production.¹⁴

There is growing evidence showing that progression of systemic inflammatory response syndrome into sepsis is due to the cellular damage and death induced by acute inflammatory responses. Cell death depends, in part, upon mitochondrial dysfunction, which is often characterized by increased production of reactive oxygen species (ROS), increased membrane permeability and eventual release of cell death mediators from mitochondria.¹⁵ Extensive mitochondrial damage leads to loss of cellular ATP pools (which can be linked to necrotic cell death) and to further inflammatory responses. Consequently, mitochondrial dysfunction contributes to a wide range of human pathologies including SIRS and sepsis. It has been suggested that therapies aimed at improving cellular redox state and energy production may be applicable for treatment of sepsis.¹⁶

The goal of this study was to show that lactoferrin modulates lipopolysaccharide (LPS)-induced changes in cellular redox state and identify potential subcellular targets of lactoferrin. Here, we report for the first time that lactoferrin protects against oxidative stress-induced mitochondrial dysfunction and DNA damage, both in cell culture and within an animal model of endotoxemia. Data derived from these studies provide an on-going mechanistic understanding by which lactoferrin decreases intensity of LPS-induced systemic inflammation. These results are also in support of lactoferrin’s potential therapeutic utility for treatment of SIRS and sepsis.

Materials and Methods

Cell culture

The AML12 cell line (non-tumorigenic parenchymal liver cells with features characteristic of matured hepatocytes) was cultured in Dulbecco’s minimal essential medium (DMEM)-Ham’s F-12 containing ITS solution and 0.1 nM dexamethasone. The U937 cell line (established from histiocytic lymphoma and displaying many monocytic characteristics) was maintained in RPMI-1640. A549 bronchial epithelial cells were cultured in Ham’s F-12 medium. All cell lines were obtained from American Type Culture Collection and were periodically tested for Mycoplasma contamination. All media (Invitrogen, Carlsbad, CA, USA) were supplemented with 2 mM L-glutamine, 1.0 mM sodium pyruvate, 100 U/ml penicillin, 100 μg/ml streptomycin (Sigma Aldrich, St Louis, MO, USA) and 10% fetal bovine serum (FBS, Atlanta Biological). The cells were routinely subcultured using trypsin-EDTA and incubated under a humidified atmosphere of 95% air and 5% CO₂ at 37°C.

Animals and treatment

BALB/c mice were purchased from Harlan Sprague-Dawley (San Diego, CA, USA). All animal experiments were performed according to the National Institutes of Health Guide for Care and Use of Experimental Animals and approved by the UTMB Animal Care and Use Committee (#0807042). The effects of lactoferrin on mitochondrial dysfunction and DNA damage induced by LPS were determined using previously established treatments and animal model.⁸ Briefly, 9–11 mice (8–10-week-old; 18–20 g per mouse) per group were injected with 5 mg lactoferrin per mouse intraperitoneally (i.p.). Twelve hours later, LPS (2.5 mg/kg body weight) was added intraperitoneally. Control groups of animals received only lactoferrin, LPS or solvent injections.

Reagents

Low endotoxin (<0.2 EU/mg by Limulus amebocyte lysate assay) human milk lactoferrin <20% iron saturated, >95% purity) was provided by PharmaReveiw Corporation (Houston, TX, USA). Rotenone (R-8875), cytochrome-c (C-3131), decylubiquinone (D7911), antimycin A (A8674), oligomycin (O-4876), pyruvate, malate, succinate, 3-nitropropionic acid, catalase were purchased from Sigma Aldrich (St Louis, MO, USA). 2′,7′-Dichlorodihydrofluorescein diacetate; dihydroethidium, MitoTracker Red and Amplex^® Red (10-acetyl-3, 7-dihydroxyphenoxazine were from Molecular Probes (Eugene, OR, USA). Lipopolysaccharide from Escherichia coli serotype O111:B4 (3 ×10⁶ EU/mg) was purchased from Sigma.

Establishment of respiration-deficient cells

Mitochondrial DNA-deficient cells were developed as described previously.¹⁷ Both cell cultures AML12 and A549 were maintained in the presence of 100 ng/ml ethidium bromide for >60 population doublings. Depletion of mitochondrial DNA (mtDNA) was confirmed by Southern blot hybridization.^17,18 Respiration-deficient cells became pyrimidine auxotrophs, and media were supplemented with uridine (50 μg/ml) and sodium pyruvate (120 μg/ml).¹⁹ For verification of the absence of mtDNA in p0 cells, DNA was isolated, treated with DNase-free RNase then digested with BamHI. After gel electrophoresis, DNA was transferred onto nitrocellulose membrane (Schleicher and Schuell BioScience, Keene, NH, USA), blocked and hybridized with a PCR-generated DNA probe for the mitochondrial genome. The forward and reverse primer sequences were: 5′-GCAGGAACAGGATGAACAGTCT-3′ and 5′-GTATCGTGAAGCACGATGTCAAGGGATGTAT-3′, respectively. The 725-bp product recognized a 10.8-kb restriction fragment when hybridized to mtDNA digested with BamHI as described previously.^17,18

Mitochondria isolation

Mitochondria were isolated from mock- and LPS-treated cells as described previously.¹⁷ Briefly, cell pellets were incubated in 10× volume of hypotonic buffer (10 mM KCl, 20 mM MOPS, and 1 mM EGTA for 20 min then Dounce-homogenized. The homogenate was centrifuged at 800 g and the supernatants re-centrifuged at 10,000 g to collect mitochondria. Mitochondrial pellets were washed, and resuspended in 10 mM KCl, 20 mM MOPS, and 1 mM EGTA containing 200 mM sucrose and 50 mM mannitol. In selected experiments, fresh mitochondrial suspensions were purified on a continuous sucrose gradient (0.25–1.5 M).

Mitochondria were also isolated from the livers of Balb/c mice. Briefly, organs of sacrificed animals were excised and rinsed in buffer A (100 mM KCl, 20 mM MOPS, 1 mM EGTA, 5 mM MgSO₄, and 1 mM ATP; pH 7.6) at 4°C. Livers were homogenized in buffer A, containing 200 mM sucrose, 50 mM mannitol, 0.2% bovine serum albumin, using a Dounce homogenizer. Isolation of mitochondria was done as described above. Fresh mitochondrial suspensions from organs were purified on a continuous sucrose gradient (0.1–1.5 M) and used immediately for determining the site(s) of superoxide anion formation or stored at −80°C for further studies.

Measurement of mitochondrial and intracellular ROS

The intracellular site of ROS generation was identified by fluorescence microscopy.^17,20 Cells were loaded with 2 μM dihydroethidium (H₂Et; Molecular Probes) for 10 min after which the cells were treated with 100 μg/ml LPS (pH 7.4) and placed in a thermo-controlled microscopic chamber. MitoTracker Red (Molecular Probes), a cell-permeable fluorescent probe that accumulates in active mitochondria, was used to stain mitochondria at a final concentration of 10 nM. Fluorescent images were captured after 60 min incubation with LPS using a Zeiss LSM510 META System driven by Metamorph^™ v.6.09 software (Universal Imaging, Downingtown, PA, USA).

A redox-sensitive probe, 2′,7′-dichlorodihydrofluorescein diacetate (H₂DCF-DA; Molecular Probes), was used to determine changes in overall cellular ROS levels.^20,21 Mock- or LPS-treated cell suspensions were loaded with 50 μM H₂DCF-DA for 15 min at 37°C. The change in fluorescence (excitation 485 nm; emission 530 nm) was measured using a FLX800 microplate reader (Bio-Tek Instruments, Winooski, VT, USA). In confirmatory studies, changes in DCF fluorescence of LPS-treated versus mock-treated cells were determined by FACSaria (Becton Dickinson, Mountain View, CA, USA). Each data point represents the mean fluorescence for 15,000 cells, from three or more independent experiments.

8-Oxo-7,8-dihydro-2′-deoxyguanosine assays

The 8-oxoG in nuclear DNA was quantified as previously described.²² Briefly, cells on microscope slides were air-dried, and fixed in acetone-methanol (1:1), rehydrated in PBS for 15 min, then sequentially treated with 100 μg/ml pepsin in 0.1 N HCl for 15–30 min at 37°C, 1.5 N HCl for 15 min, and sodium borate for 5 min. The cells were incubated with non-immune IgG (100 μg/ml) for 30 min and washed in PBS containing 0.5% bovine serum albumin, 0.1% Tween 20 (PBS-T). After incubation with anti-8-oxoG antibody (Trevigen, Gaithersburg, MD, USA; 1:200 dilutions)²³ for 30 min, the cells were washed three times with PBS-T for 15 min then exposed to fluorescein-conjugated secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 60 min. Cells were washed with PBS-T for 15 min (3-times) and their DNA stained with DAPI (10 ng/ml) for 15 min. The cells were air dried and mounted in anti-fade medium (Dako North America, Inc., Carpinteria, CA, USA) on a microscope slide. The fluorescence intensities of a minimum of 40 cells per plate were determined using a Zeiss LSM510 META system, operated via MetaMorph software v.6.06r (Universal Imaging Corporation).

Oxygen consumption

The oxygen consumption rates of mitochondria were determined at 30°C with a Clark-type oxygen electrode (Strathkelvin Oxygen System Model 782, Strathkelvin Instruments, UK) as previously described.^17,24 Purified mitochondria (0.2 mg) were suspended in 1 ml respiration medium (125 mM KCl, 20 mM HEPES, pH 7.4), 5 mM potassium phosphate, ±substrates (5 mM pyruvate plus 5 mM malate as well as 10 mM succinate). The signal from the oxygen sensor was recorded on a computer at sampling intervals of 0.5 s with the aid of software from Strathkelvin Instruments (782 System v.3.0). Respiration was measured without ADP (state IV) and with 0.5 mM ADP (state III).¹⁷ Mitochondrial suspensions showing higher than 3 respiratory control ratios were used in Amplex Red assays.

Amplex Red assay

Amplex Red^® (10-acetyl-3,7-dihydroxyphenoxazine; Molecular Probes) reacts with H₂O₂ in the presence of horseradish peroxidase (HRP) to generate a stable product, resorufin.²⁵ Briefly, mitochondria (100 μg/ml) were suspended in 100 μl (per well) reaction buffer and incubated at room temperature (25°C) for 30 min with 0.25 U/ml (determined in preliminary studies) of Amplex Red and 0.5 U/ml of HRP.^17,24 The increase in fluorescence (with excitation and emission wavelengths of 563 and 587 nm, respectively) was measured using a microplate reader (SpectraMass M2, Molecular Devices Inc.). The rate of H₂O₂ production was linear with mitochondrial protein concentration. Reactions were carried out with exogenously added superoxide dismutase (SOD). The addition of catalase (400 U/ml, Sigma) decreased H₂O₂ levels by ~90%. As a positive control, increasing concentrations of H₂O₂ (0–400 pmol) were used.

Estimation of mitochondrial and nuclear DNA damage

About 3 ×10⁶ cells (U937) were plated and treated with LPS (100 ng/ml). DNA was extracted using a genomic DNA extraction kit (Qiagen, Chatsworth, VA, USA) using the protocol supplied with the kit. The concentration of DNA was determined using the PicoGreen^® dsDNA Quantitation Kit (Molecular Probes). Free PicoGreen dye is essentially non-fluorescent and exhibits>1000-fold fluorescence enhancement upon binding to dsDNA at an excitation and emission wavelength of 480 and 530, respectively. The assay displays a linear correlation between dsDNA quantity and fluorescence, being extremely sensitive (detection range extending from 25 pg/ml to 1 μg/ml). Quantitative PCR was performed using a protocol described previously,^26,27 and the quantitation of PCR products was done using PicoGreen dye as previously described.²⁸

The primer nucleotide sequences for U937 were: for the 17.7-kb 5′-flanking region of the β-globin gene 5′-TTGAGACGCATGAGACGTGCAG-3′ (forward), and 5′-GCACTGGCTTAGGAGTTGGACT-3′ (reverse) and for the 16.2-kb fragment of the mitochondrial genome, 5′-TGAGGCCAAATATCATTCTGAGGGGC-3′ (forward) and 5′-TTTCATCATGCGGAGATGTTGGA TGG-3′ (reverse).²⁹

The primer nucleotide sequences for AML12 were: for the 7.2-kb fragment of the β-globin gene 5′-GGAGCAAGGTCCAGGGTGAAGAA-3′ (forward) and 5′-TTTGCATCCAGATCATGGTCCCT-3′ (reverse) and for the 10.4-kb mitochondrial fragment 5′-GCCAGCCTGACCCATAGCCATAATAT-3′ (forward) and 5′-GATGGTTTGGGAGATTGGTTGAT GT-3′ (reverse).³⁰ The PCR was initiated with a 75°C hot-start addition of the polymerase and allowed to undergo the following thermocycler profile: an initial denaturation for 1 min at 94°C followed by 25 cycles of 94°C denaturation for 15 s and 68°C primer extension for 12 min. A final extension at 72°C was performed for 10 min at the completion of the profile. To ensure quantitative conditions, a control reaction containing 7.5 ng of template DNA was included in amplification reactions. For quality control and specificity of primers, an aliquot of each PCR product was resolved on a 1% agarose gel and electrophoresed in TBE (90 mM Tris, 64.6 mM boric acid, 2.5 mM EDTA, pH 8.3) at 80 V (5 V/cm) for 4 h. DNA lesion frequencies were calculated as described previously.^28,31 Briefly, the amplification of damaged samples (A_D) was normalized to the amplification of a non-damaged control (A_O), resulting in a relative amplification ratio. Assuming a random distribution of lesions and using the Poisson equation (fx = e^−λλ^x/x, where λ is the average lesion frequency for the non-damaged template, i.e. the zero class x = 0), the average lesion per DNA strand was determined as λ = −lnA_D/A₀.

Statistical analysis

Results were analyzed for significant differences using ANOVA procedures and Student’s t-tests (Sigma Plot v.6.0). Data are expressed as the mean ± SE. Results were considered significant at P<0.05. (*P<0.05, **P<0.01, ***P<0.001, ****P<0.0001).

Results

Effect of lactoferrin on LPS-induced cellular ROS levels

To investigate the changes in cellular ROS levels, U937 (and AML12) cells were treated with increasing concentrations of LPS and 45 min later cells were loaded with H₂DCF-DA (50 μM, for 15 min). The changes in LPS-induced ROS levels were dose-dependent from 25 ng/ml to 100 ng/ml; higher doses (200, 400 and 800 ng/ml) did not further increase fluorescent levels (Fig. 1A). Similar results were obtained using both U937 (Fig. 1A) and AML12 cells (Fig. 1A, inset). In additional studies, LPS was used at a concentration of 100 ng/ml. As shown in Figure 1B, LPS induced a biphasic increase in ROS levels. The first wave of ROS was observed at ~1 h and another one was seen ~6 h after LPS addition (Fig. 1B). Treatment of cells with H₂O₂ (50 μM) or glucose oxidase (GO; 100 ng/ml), an enzymatic oxidative stress generator,²⁰ resulted in a single wave change in DCF fluorescence (Fig. 1B, inset).

Fig. 1.

Dose-dependent changes in LPS-induced ROS levels. (A) Parallel cultures of cells were treated with increasing concentrations of LPS and 45 min later cells were loaded with H₂DCF-DA (50 μM, 15 min). (B) Cells were treated with LPS (100 ng/ml) for increasing time periods as indicated. Inset: changes in ROS levels by H₂O₂ (50 μM, in serum-free medium) or glucose oxidase (GO, 100 ng/ml) an enzymatic ROS generator. Filled squares, GO; filled diamonds, H₂O₂; filled triangle, mock-treated. At 15 min before assessment of ROS levels, cells were loaded with H₂DCF-DA (50 μM). The changes in fluorescence intensities were assessed in an FLx800 microplate reader (Bio-Tek Instruments) at 488 nm excitation and 530 nm emission. Results are mean ±SEM (n = 4–7).

There are a number of hypotheses for the antioxidant properties of lactoferrin, the most obvious one is linked to its iron sequestering capability.³² To test lactoferrin’s antioxidant potential, cells (U937, AML12) were pre-treated for 3 h with increasing concentrations of lacto-ferrin, and 100 ng/ml LPS was then added. Lactoferrin, beginning at 8 μg/ml and at higher concentrations, significantly (P<0.05) decreased LPS-induced ROS levels; 64 μg/ml and 128 μg/ml were the most effective doses for use in U937 cells (Fig. 2A). Similar results were obtained for AML12 cells (data not shown). In controls, 64 μg/ml lactoferrin also decreased H₂O₂- and glucose oxidase-induced increases in ROS levels in both U937 (not shown) and AML12 (Fig. 2A, inset). Deferroxamine (DFO), an iron sequestering compound,³³ decreased LPS-induced ROS levels, but not significantly (Fig. 2B). Substantial decrease by DFO in ROS levels was obtained after H₂O₂ and glucose oxidase treatment of cells as shown previously.¹⁴ The level of DCF-mediated fluorescence was partially decreased when cells were pre-incubated with diphenyleneiodonium (DPI; Fig. 2B), an NADPH oxidase inhibitor.³⁴ To test if LPS-induced increase in DCF signal was specific for ROS, antioxidants were used. It is noteworthy that 64 μg/ml lactoferrin decreased ROS-mediated DCF fluorescence similarly to N-acetyl-L-cysteine (NAC, 10 mM; Fig. 2A,B), an antioxidant and a free radical-scavenging agent that increases intracellular GSH,³⁵ and Ebselen (10 μg/ml), a thioredoxin reductase-dependent catalyst for α-tocopherol quinone reduction in the mitochondria.³⁶

Fig. 2.

Lactoferrin decreased LPS-induced ROS levels in a dose-dependent manner. (A) Parallel cultures of U937 cells were treated for 3 h with increasing concentrations of lactoferrin as indicated. As control, cells were treated with 10 mM N-acetyl-L-cysteine (NAC) for 3 h. LPS (100 ng/ml) was added for 45 min and cells were loaded with H₂DCF-DA (50 μM, 15 min). Inset: AML12 cells were treated with 64 μg/ml lactoferrin and H₂O₂ or glucose oxidase (GO) was added. (B) Parallel cultures of U937 and AML12 cells were treated with lactoferrin (64 μg/ml), NAC, (10 mM), DFO (64 μg/ml) or Ebselen (10 μM) for 3 h and LPS was added for 45 min. The changes in fluorescence intensities were assessed in an FLx800 microplate reader (Bio-Tek Instruments) at 488 nm excitation and 530 nm emission. Results are mean ±SEM (n = 4–7 independent experiments). **P<0.01, ***P<0.001. LF, lactoferrin; NAC, N-acetyl-L-cysteine; GO, glucose oxidase; DFO, deferoxamine; DPI, diphenyleneiodonium.

Lipopolysaccharide increases cellular ROS levels via mitochondria

Lipopolysaccharide increases intracellular levels of ROS, via NADPH oxidases in macrophages and mitochondria in neutrophils;^37,38 however, there are no reports for these effects in monocytic cells (U937) or hepatocytes (AML12). To identify the intracellular site of ROS production, AML12 cells were LPS-treated at 50% confluence, and then loaded 30 min later with dihydroethidium (2 μM) for 15 min; changes in fluorescence were recorded by microscopy. The green fluorescence mediated by dihydroethidium +superoxide reaction products³⁹ co-localized with MitoTracker Red suggesting that the mitochondria are the primary sites of ROS generation (Fig. 3A). Intensity of fluorescence by dihydroethidium + superoxide reaction products in lactoferrin or NAC pre-treated LPS-exposed cells was decreased nearly to levels seen in mock-treated controls cells. Similar fluorescence intensity of mitochondria-mediated dihydroethidium + superoxide reaction products was observed in U937 cells (data not shown).

Fig. 3.

LPS increases ROS levels via mitochondria. (A) Parallel cover-slip cultures of AML12 cells were treated with lactoferrin (64 μg/ml) or NAC (10 mM) for 3 h and LPS (100 ng/ml) was added. After 45 min, cells were loaded with MitoTracker red (10 nM) and dihydroethidium (2 μM) for 10 min. Fluorescent images were captured after 60 min incubation with LPS using a Zeiss LSM510 META System driven by Metamorph v.6.09 software. (B) mtDNA-depleted (p0 cells) and wild-type cells were treated with LPS (100 ng/ml) and 45 min later cells were loaded with H₂DCF-DA (50 μM, for 15 min). Inset: p0AML12 cells were treated with H₂O₂ or LPS ± DPI (10 μM) and 1 h later changes in DCF fluorescence was assessed. Changes in fluorescence intensities were assessed in an FLx800 microplate reader.²⁴ Results are mean ± SEM (n = 4–7 independent experiments). DPI, diphenyleneiodonium.

To test further whether LPS increases cellular ROS levels from mitochondria, we utilized mtDNA depleted p0AML12 and p0U937 cells.¹⁷ When p0 cells were exposed to LPS, increase in ROS levels were significantly (P<0.05) lower compared to wild-type cells (Fig. 3B). In control experiments, addition of H₂O₂ to p0 cells increased DCF-signals similar to wild-type cells. Addition of DPI, nearly abolished LPS-induced increase in ROS levels in mtDNA depleted cells (Fig. 3B, inset). On the other hand, DPI only partially inhibited LPS-induced ROS generation in wild-type cells (Fig. 2B). Together, these data suggest that mitochondria are the primary source of ROS generation in LPS-treated cells.

Decreased mitochondrial release of H₂O₂ by lactoferrin

Next, mitochondria were isolated from ±LPS-exposed cells to determine the amount of organelle-specific H₂O₂ (originating from O₂^•− dismutation by superoxide dismutase) release. First, lactoferrin was investigated to determine its effect on LPS-induced release of H₂O₂ and, second, experiments were performed to define the site of ROS generation in the mitochondrial respiratory chain. Mitochondria from LPS-treated AML12 (and U937 cells; data not shown) cells release approximately three times more H₂O₂ (compared to mitochondria from mock-treated cells; Fig. 4A). When LPS-exposed cultures were pretreated with lactoferrin, mitochondrial release of H₂O₂ was not significantly different from those of mock-treated cells (Fig. 4A). Deferoxamine did not significantly decrease H₂O₂ levels in mitochondria isolated from LPS-treated cells (Fig. 4A).

Fig. 4.

Lactoferrin decreases LPS-induced mitochondrial release of H₂O₂. (A) Cells (AML12) were propagated to large quantities, ±lactoferrin (64 μg/ml), ±DFO (64 μg/ml) or ±NAC (10 mM) for 3 h and then LPS-challenged for 1 h. Cells were harvested and collected by centrifugation, homogenized to isolate and purify mitochondria. Hydrogen peroxide release during 30 min incubation period from mitochondrial suspension (100 μg protein/ml/sample) respiring pyruvate + malate + succinate were determined by Amplex red assays (see Materials and Methods). Results are expressed as mean ±SEM values of at least three independent experiments (n = 4 or 5) ***P = 0.001. (B) Tentative site(s) of superoxide anion formation in mitochondrial respiratory chain was identified by use of inhibitors of electron transport. Mitochondria (100 μg protein/ml/sample; respiring pyruvate + malate + succinate) from LPS-treated cells were isolated, purified and the released H₂O₂ (during 30 min) in the presence or absence of inhibitors was determined. Rot, rotenone (10 μM); 3-NPA, 3-nitropropionic acid (1 mM); AA; Antimycin A (3 μM); Stigm, stigmathellin (0.6 μM). Inhibitory concentrations were determined in preliminary studies (not shown). In (B), the addition of catalase to mitochondrial suspension decreased H₂O₂ levels by ~95%. Results are expressed as mean ± SEM values of at least three independent experiments (n = 3–5), ***P = 0.001. LF, lactoferrin; NAC, N-acetyl-L-cysteine; DFO, deferoxamine.

To identify the site of ROS generation, specific inhibitors of the respiratory complexes were applied to mitochondrial suspensions respiring pyruvate/malate plus succinate (complex I and II substrates, respectively). Both rotenone (an inhibitor of NADH–ubiquinone reductase activity of complex I) and 3-NPA (an inhibitor of succinate dehydrogenase) individually decreased LPS-induced H₂O₂ release and together they lowered it nearly to background levels (Fig. 4B). Antimycin A (an inhibitor of cytochrome-b re-oxidation in complex III) significantly (P<0.05) increased H₂O₂ release when added to mitochondria from LPS-exposed cells. Stigmatellin, an inhibitor of entry of electrons into complex III, inhibited H₂O₂ production from mitochondria isolated from LPS-exposed cells, when used at a concentration of 0.6 μM. Stigmatellin also showed inhibition when antimycin A was added to mitochondria from LPS-exposed cells (Fig. 4B). Addition of catalase decreased H₂O₂ to an undetectable level under our assay conditions (Fig. 4B). Although additional studies are required to identify precisely the primary site of LPS-induced mitochondrial ROS generation, these result indicate that respiratory complex III is the major site of superoxide anion generation.

To obtain information if mitochondria-generated ROS can diffuse into other subcellular compartments, such as the nucleus, 8-oxo-7,8-dihydro-2′-deoxyguanosine (8-oxoG) levels were determined in nuclear DNA as a marker for oxidative cellular stress.⁴⁰ AML12 cells were pretreated with lactoferrin (for 3 h) and subjected to oxidative stress by addition of LPS or H₂O₂. Data showing that LPS, in a similar manner to H₂O₂, increases 8-oxoG levels in nuclear DNA (Fig. 5) in line with their pro-oxidant capacity shown in DCF assays. Importantly, in the presence of lactoferrin, a significant (P<0.05) decrease in 8-oxoG levels was observed. As a control, NAC (3-h pretreatment prior to LPS or H₂O₂ addition) decreased 8-oxoG formation (Fig. 5), and DFO significantly decreased 8-oxoG levels only in H₂O₂-treated cells (Fig. 5). These data provide additional support for the specific nature of LPS-mediated ROS generation and help to clarify antioxidant properties of lactoferrin.

Fig. 5.

Lactoferrin prevents LPS-increased 8-oxoG levels. Parallel cover-slip cultures of AML12 cells were treated for 3 h with lactoferrin (64 μg/ml) and 100 ng/ml LPS was added for 1 h. In controls, the cells were pretreated with NAC (10 mM) or DFO (64 μg/ml) and then exposed to LPS (100 ng/ml) for 1 h. Similar studies were undertaken using 50 μM H₂O₂ (in serum-free medium) as controls. After treatment cells were washed, dried and fixed in acetone-methanol for 10 min and processed for staining with antibody to 8-oxoG (as in Material and Methods). After staining, cells were air dried and mounted in anti-fade medium. The fluorescence intensities of a minimum of 40 cells per sample were determined using a Zeiss LSM510 META system, operated via MetaMorph software v.6.06r. Results are expressed as mean ±SEM values of at least three independent experiments (n = 4 or 5) **P>0.01, ***P = 0.001. LF, lactoferrin; NAC, N-acetyl-L-cysteine; DFO, deferoxamine.

Lactoferrin attenuates mitochondrial dysfunction in liver of LPS-treated mice

Next, we sought to determine whether the observed protection from mitochondrial dysfunction by lactoferrin occurs in LPS-treated animals. Balb/c mice were lactoferrin (or mock)-treated (for 12 h) and then injected with LPS as previously described.¹⁰ Twelve hours later, livers were excised and mitochondria were isolated. As shown in Figure 6A, mitochondria from LPS-treated animals released significantly (P<0.05) higher amounts of H₂O₂ than those isolated from lactoferrin only or saline-injected animals. Importantly, mitochondria from lactoferrin-pretreated plus LPS-challenged animals produced significantly (P<0.05) less H₂O₂. In addition, mitochondria isolated from heart, skeletal muscle, and brain of lactoferrin-pretreated LPS-treated mice released significantly (P<0.05) lower amounts of H₂O₂ compared to mitochondria isolated from LPS-challenged mice in the absence of lactoferrin pretreatment (Fig. 6B). When animals were injected with DFO (5 mg per mouse), only partial reduction in H₂O₂ levels was observed from the mitochondria isolated from LPS-challenged animals (Fig. 6A). On the other hand, NAC significantly inhibited mitochondrial H₂O₂ release (P^<0.05). Together, these results strongly suggest that lactoferrin attenuates mitochondrial dysfunction in vivo and support the hypothesis that cellular effects of lactoferrin are more complex than those based on iron sequestration.

Fig. 6.

Lactoferrin decreases mitochondrial H₂O₂ release in LPS-treated mice. (A) Balb/c mice were lactoferrin- (5 mg per mouse) (or mock)-treated (for 12 h) and then injected with LPS (2.5 mg/kg) as in Materials and Methods. After 12 h, livers were excised and mitochondria were isolated. In controls, animals were pretreated with NAC (50 mg/kg) and DFO (5 mg per mouse) intraperitoneally then LPS-treated. Livers were excised and mitochondria were isolated. H₂O₂ released from purified mitochondria was determined by Amplex Red assay (Materials and Methods). Results are expressed as mean ± SEM values from 6–9 animals in independent experiments, **P>0.01, ***P = 0.001. (B) Decreased release of H₂O₂ from mitochondria isolated from heart, skeletal muscle and brain lactoferrin-treated mice. Animals were pretreated with lactoferrin or saline and LPS-challenged. Heart, skeletal muscle, brains and lives were collected and release of H₂O₂ from purified mitochondria determined. Results are expressed as mean ±SEM from 4–6 animals, **P>0.01, ***P = 0.001. LF, lactoferrin; NAC, N-acetyl-L-cysteine; DFO, deferoxamine.

Low levels of LPS-mediated oxidative damage to mtDNA in the presence of lactoferrin

The release of H₂O₂ to the extramitochondrial space is a good indication for mitochondrial dysfunction; however, it does not provide evidence for release of ROS into the mitochondrial matrix where mtDNA resides. Therefore, the changes in mtDNA damage levels were further investigated by quantitative PCR, utilizing AML12 (and U937) cells treated with LPS. As shown in Figure 7A, the PCR-amplifiable levels of mtDNA decreased significantly (P<0.05) in LPS-treated U937 cells. Damage to a nuclear (β-globin) gene was used as control. Lipopolysaccharide-treatment significantly (P<0.05) increased levels of both mtDNA and nuclear DNA damage. Time-course experiments showed that mtDNA damage persisted for a longer time period compared to damage to control β-globin gene sequences (Fig. 7A). Persistence of mtDNA damage is consistent with sustained ROS generation, as shown in the DCF assays above (Fig. 1B). Remarkably, levels of damaged DNAs were less in lactoferrin-pretreated cells. While NAC significantly (P<0.05) lowered damage to DNA, DFO pretreatment of cells had detectable, but not significant, effects on mtDNA damage after LPS addition (Fig. 7B).

Fig. 7.

Lactoferrin decreases mitochondrial DNA damage after LPS challenge. (A) Kinetic changes in LPS-induced damage to mitochondrial and nuclear DNA. Parallel cultures of AML12 cells were treated with LPS (100 ng/ml) for 60 min and genomic DNAs were isolated at 0, 1, 2, 4, 6, 12, and 18 h. DNA concentrations were determined using PicoGreen (dsDNA Quantitation Kit) as in Material and Methods. Q-PCR was undertaken using primer pairs for mitochondrial DNA (16 kb) and the nuclear β-globin DNA (17 kb) as described previously.²⁶,²⁷ The quantification of PCR products was determined by fluorescence intensities of PicoGreen-mediated fluorescence as previously described.²⁸ Results are expressed as fluorescence intensities ± SEM from 3–5 independent experiments. (B) Parallel cultures of AML12 cells were pretreated with lactoferrin (64 μg/ml), DFO (64 μg/ml) and NAC (10 mM) for 3 h and LPS (100 ng/ml) was added for 60 min. Genomic DNAs were isolated (at 2 h and 6 h after LPS addition) and DNA concentrations, levels of intact mitochondrial DNA were determined. Results are expressed as fluorescence intensities ±SEM from 3 or 4 independent experiments. ***P<0.001; n.s, not significant. (C) Mice (6–9 per group) were pretreated with lactoferrin, DFO, or NAC for 3 h and intraperitoneally injected with LPS (2.5 mg/kg). Livers were excised 12 h after LPS challenge and genomic DNAs were quantified (as above). Levels of intact mitochondrial DNAs were determined by quantitative PCR. In (A), (B) and (C) thermocycler profile: an initial denaturation for 1 min at 94°C followed by 25 cycles of 94°C denaturation for 15 s and 68°C primer extension for 12 min. The quantification of PCR products was done using PicoGreen dye (Materials and Methods). Results are expressed as fluorescence intensities ± SEM from PCRs of 6–9 livers per treatment. **P<0.01; ***P<0.001. LF, lactoferrin; NAC, N-acetyl-L-cysteine; DFO, deferoxamine.

Next, the changes in levels of mtDNA damage were determined in livers of animals treated with LPS and/or lactoferrin. As shown in Figure 7C, at 12 h after LPS administration, the levels of undamaged mtDNA decreased in the liver compared to time 0. These results are in line with levels of released H₂O₂ (Fig. 4). Lactoferrin, DFO, or NAC alone did not change mtDNA levels during the pretreatment period (Fig. 7C); thus, it was similar to saline-injected control animals (data not shown). However, quite strikingly, lactoferrin was able to increase levels of intact (non-damaged) mtDNA in the LPS-treated mice (Fig. 7C). While NAC showed significant protection, treatment of mice with DFO only partially decreased mtDNA damage after LPS challenge (Fig. 7C). Together, these result show that lactoferrin decreases mitochondrial ROS generation, and has the potential to lower LPS-mediated damage to mtDNA. Because DFO did not provide significant protection, this suggests that iron sequestration itself is not the major mechanism by which lactoferrin protects mtDNA. These data also imply that ROS generated by the respiratory complexes are released both to the extramitochondrial space and mitochondrial matrix, where they can overwhelm existing antioxidant machineries.

Discussion

Lactoferrin is a natural immunomodulator and antimicrobial agent. It is also known to reduce toxic effects of LPS by both direct and indirect mechanisms in vivo. As such, lactoferrin is considered for therapeutic use in systemic inflammatory response syndromes (reviewed by Caccavo et al.⁴). Data supporting the potential clinical utility of lactoferrin are overwhelming; however, the exact cellular mechanism(s) by which lactoferrin demonstrates preventive and/or therapeutic potential is not yet known. Here, we demonstrate that lactoferrin decreases intracellular oxidative stress levels induced by LPS via reduced mitochondrial dysfunction in cultured cells. Importantly, lactoferrin attenuated mitochondrial dysfunction in the liver and other organs of LPS-treated animals, as demonstrated by reduced release of H₂O₂ from mitochondria and significantly lower mitochondrial DNA damage. In contrast, deferoxamine, an iron chelator, provided only marginal protection of mitochondria after LPS treatment. Thus, lactoferrin protects against oxidative insults at the cellular level via a more complex mechanism than simple iron sequestration.

There is growing evidence to show that progression of systemic inflammatory response syndrome into sepsis is due to cellular damage and death induced by acute inflammatory responses. Both apoptotic and necrotic cell death are tightly associated with mitochondrial dysfunction, often characterized by increased production of ROS, increased membrane permeability, loss of mitochondrion integrity and alterations in cellular ATP levels.¹⁵ Mitochondrial dysfunction and attendant bio-energetic defects are indeed increasingly becoming recognized as important role players in both chronic and acute disorders, including sepsis.⁴¹ The results presented here reveal that pretreatment of cultured cells and animals with lactoferrin led to decreased LPS-induced elevation in ROS levels, lessened damage to nuclear and mtDNA, and overall sustained protection against mitochondrial dysfunction. These findings are supported by previous reports demonstrating protective capabilities of lactoferrin in experimental mice treated with lethal doses of bacteria or LPS.⁹ In these studies, prophylactic or therapeutic administration of lactoferrin significantly decreased levels of pro-inflammatory mediators (TNF-α, IL-6 and IL-10) and nitric oxide levels after LPS injection.¹⁰

Evidence supports that LPS-induced mitochondrial dysfunction is linked to mtDNA depletion and results in inhibition of mitochondrial transcription, although the precise molecular mechanism for this function has not yet been clearly established.^38,42,43 In our studies, cultured U937 (cells of monocytic origin) and AML12 cells (immortalized hepatocytes) were used as in vitro links to identify potential liver-related dysfunction. The latter was important relative to findings identified in livers of experimental animals. In vitro, a consistent biphasic increase in ROS levels was observed. We propose that the second ROS wave is a consequence of initial oxidative insult/damage to mitochondrial respiratory complex proteins and mtDNA after LPS exposure. Ebselen, an antioxidant that exerts a primary effect within the mitochondria,³⁶ decreased the level of initial oxidative insult, and LPS did not induce a second stress wave although it remained present in the medium. Data generated by the use of mitochondrial DNA depleted p0 cells further support the mitochondrial origin of cellular ROS. mtDNA depleted p0 cells maintain mitochondrial membrane potential via utilization of ATP generated by glycolysis to generate low levels of ROS;^44,45 thus, these cells respond poorly to agents that induce mitochondrial ROS via receptor-mediated signal transduction pathways (such as LPS). Indeed, in our p0 cells, LPS-induced ROS levels were significantly lower than in wild-type cells, indicating that the majority of ROS are generated via mitochondria after LPS exposure.

In an attempt to identify the mitochondrial site of ROS generation, it was shown here that the initial oxidative insult did not have detectable effect on the function of respiratory complexes I and II and electron transport to the ubiquinol pool. Specifically, rotenone (inhibits ubiquinone reduction site of complex I, while its NADH oxidoreductase activity remains unaffected⁴⁶) and 3-NPA (a succinate dehydrogenase inhibitor⁴⁷) individually only partially decreased superoxide levels, while together they lowered release of H₂O₂ from mitochondria to nearly background levels. These observations imply that electron flow into respiratory complex III is required for LPS-induced ROS generation. In support, stigmatellin (which acts at the Qo center of the cytochrome-bc₁ complex, binds to the heme b-566 domain of cytochrome-b as well as to the Fe₂–S₂ protein in respiratory complex III),⁴⁸ alone significantly lowered H₂O₂ release from mitochondria in LPS-treated cells. Antimycin A (which binds to the matrix side of complex III and inhibits the Qi site of cytochrome-c oxidoreductase, in the cytochrome-b subunit⁴⁹) alone increased mitochondrial H₂O₂ release; however, when it was added to mitochondria from LPS-treated cells, the H₂O₂ concentrations were synergistically increased. Although it requires further investigation, these data together imply that the mitochondrial site of ROS release occurs at complex III in LPS-exposed cells. Indeed, complex III is one of the major sources of electrons that reduce molecular oxygen to superoxide anion in pathophysiological conditions.⁵⁰ Previous studies have suggested that inhibition of complex I by rotenone increased intracellular levels of ROS in neutrophils in response to LPS.⁵¹ Another study showed that LPS-induced activation of redox-sensitive MAPK occurs via the mitochondrial uncoupling protein, UCP2 in macrophages.⁵² These observations are in line with our data while it is possible that there are cell-type specific variations in responses to LPS. Nonetheless, further studies are in progress to define the precise mitochondrial site of ROS generation and its biological significance.

To investigate if mitochondrial ROS change the redox conditions in the nucleus and also in the mitochondrial matrix, specific studies were undertaken to determine oxidative DNA damage levels in subcellular compartments. Among DNA bases, guanine is the most susceptible to oxidation, forming 8-oxoG, a two electron oxidation product found in the RNA, DNA and GTP pool of prokaryotic and eukaryotic cells.⁴⁰ Significant increases in 8-oxoG levels in nuclear DNA were demonstrated, suggesting that mitochondrial superoxide anion was converted via enzymatic and/or iron-mediated dismutation into H₂O₂ which diffused into the nuclei of cells. In the nucleus, directly or via iron-mediated dismutation, hydroxyl radicals (most reactive species) are formed from H₂O₂ that cause damage to DNA bases, such as guanine, which has the lowest redox potential among the DNA bases. Most importantly, lactoferrin lowered 8-oxoG levels to nearly background levels in LPS-treated cells.

Changes in levels of mtDNA damage were determined by the well-validated quantitative PCR method.^27–29 The increase in damage to mtDNA provides evidence that LPS-induced ROS overwhelms antioxidant capacity of the mitochondrial matrix. Time-course analysis showed that mtDNA damage persisted for a substantial longer period of time, compared to accumulated nuclear DNA damage. Mitochondria have efficient and abundant machineries for oxidative DNA damage repair,⁵³ therefore, presence of persisting mtDNA damage as a consequence of increased mitochondrial ROS generation after LPS treatment has yet to be defined. We propose that excessive damage to mitochondrial respiratory chain proteins and mtDNA by the initial oxidative stress is a cause for increased ROS levels, which feed back and induce additional damage to DNA and proteins that sustain ROS generation both in cultured cells or LPS-challenged animals. Indeed, after a single intraperitoneal injection of LPS, liver mtDNA copy number decreased, as determined by Southern analysis, within 24 h.⁴² This provides new insight into the physiological significance of mtDNA mutagenesis and perhaps into natural mechanisms to offset the pathological consequences of environmental insults.

An important question of how lactoferrin may protect against mitochondrial dysfunctions has been addressed in both cultured cells and in mice treated with LPS. A possibility is that lactoferrin attenuates cellular responses to LPS exposure which may involve iron binding, while yet undefined events may not be excluded. DFO, a naturally occurring iron-chelating agent, has been successfully used for clinical indications of iron overload.^54,55 In our studies, we compared the effects of DFO to lactoferrin on preventing mitochondrial dysfunction (release of H₂O₂, mtDNA damage in cultured cells and liver). DFO decreased the levels of released H₂O₂ from mitochondria and lessened mtDNA damage, but not significantly, indicating that iron chelation itself is not sufficient for prevention of LPS-induced mitochondrial dysfunction. In line with our data, lactoferrin decreased cellular oxidative stress levels and airway inflammation, while the DFO effect was insignificant both in cultured cells and in a mouse model of allergic inflammation.¹⁴

Together, these data suggest that iron chelation by lactoferrin is not the only explanation for its beneficial effects on mitochondrial function. In support, our transcriptomal network analysis showed regulated changes at RNA levels during cellular response to lactoferrin-exposure (unpublished data by Boldogh and colleagues). Finally, these data also support the significance of lactoferrin in the resolution or progression of the immune responses during the development of insult-induced metabolic imbalance, including the initial reactions to infectious assault, trauma, and injury. These findings may be critically important in the development of clinical protocols to limit pathological damage by sepsis.

Conclusions

Systemic inflammatory response syndrome is a progressive and life-threatening condition that remains a cause of high mortality. The majority of current treatments against sepsis fail to demonstrate significant clinical benefit. Our results support use of lactoferrin in prevention and therapy, as a direct therapeutic agent to combat ensuing mitochondrial dysfunction and generation of ROS that culminates in ultrastructural mitochondrial abnormalities and signals for cell destruction within the affected tissue.^56,57 Overall, lactoferrin has a potential to protect against a cellular-based organ failure in systemic inflammatory response syndrome and sepsis.^42,43,58

Abbreviations

3-NPA

3-nitropropionic acid

H₂Et

dihydroethidium

DFO

deferoxamine

LPS

lipopolysaccharide

O₂^•−

superoxide anion

ROS

reactive oxygen species

SDH

succinate dehydrogenase

SOD

superoxide dismutase