Heme-induced heme oxygenase-1 (HO-1) in humanmonocytes inhibits apoptosis despite caspase-3up-regulation

Detlef Lang1, Stefan Reuter1, Tania Buzescu1, Christian August2 and Stefan Heidenreich1

1Department of Medicine D and 2Institute of Pathology, University of Muenster, Muenster, Germany

Keywords: apoptosis, caspase-3, heme oxygenase-1, monocytes

Abstract

Monocyte activation, apoptosis and differentiation are hallmarks of most inflammatory vasculardisorders. We studied the effects of heme oxygenase-1 (HO-1) induced by its substrate hemin onapoptosis, caspase-3 expression and the differentiation of freshly isolated human monocytes. Hemininduced HO-1 in a dose- and time-dependent fashion as measured by semi-quantitative RT–PCR andflow cytometry. Apoptosis was markedly suppressed by hemin in cells rendered apoptotic by serumdeprivation or dexamethasone as determined by flow cytometric detection of annexin V binding ortransmission electron microscopy (TEM). The specific HO-1 inhibitor zinc protoporphyrin (ZnPP)reversed the effects of hemin on monocyte apoptosis and diminished cell lifespan. Surprisingly, thecytoprotective effects of hemin were positively correlated with caspase-3 up-regulation. Hemin-induced apoptosis suppression was enhanced by the caspase-3 inhibitor DEVD-CHO, indicating thatcaspase-3 was active in a pro-apoptotic fashion. Hemin inhibited CD95 as a putative cytoprotectivemechanism. Morphological studies and detection of CD86 showed that monocytes differentiated intomacrophages in response to hemin after relatively long incubation times, a phenomenon that mightbe provoked by caspase-3-regulated pathways. Our results confirm a similar cytoprotective effect ofhemin/HO-1 for monocytes as has been shown for other cells, despite caspase-3 up-regulation. Thefact that HO-1 may adversely affect monocyte survival and differentiation could be of particularsignificance in future therapies for occlusive vascular diseases or transplant rejection.

Introduction

Heme oxygenase-1 (HO-1) is one of three enzyme isoformswith the capacity to degrade tissue heme (1,2). Heme oxy-genases (HO) are able to open the heme ring and convertheme into biliverdin, triggering the release of iron and carbonmonoxide. This is the initial and rate-limiting step in hemedegradation. HO-1 is the inducible HO isoform that respondsrapidly to diverse stimuli (3) and protects tissues againsta wide range of injuries including nephrotoxic (4), hypoxic (5),hyperoxic (6), ischemic (7) and inflammatory (8) disorders.These protective effects of HO-1 result from enzyme actionsthat block injury pathways by, for example, reducing hemeexpression, a pro-oxidant species (9) or by generating carbonmonoxide, a gaseous product that is vasodilatory, anti-apoptotic and anti-inflammatory (6,10). In addition, biliverdinand bilirubin have antioxidant and cytoprotective effects thatmay enhance HO-1 effects (11).

The significance of HO-1 expression for protection againstinjury to and counter-regulation of apoptotic pathways has notbeen elucidated for freshly isolated human monocytes. Mono-cytes/macrophages play a central role in the host’s innate andacquired immunity by providing protection against infectiousmicro-organisms and tumours. Monocytes in a culture envir-onment can be rescued from constitutive occurring apoptosisthrough serum supplementation or the addition of growthfactors (12,13). The survival or apoptosis of monocytic cellscan be crucial factors in determining whether an inflammatorydisease spreads or subsides. Gaining insight into the mech-anisms involved in the induction or inhibition of apoptoticevents could also have therapeutic implications. The presentstudy investigated the role of hemin-induced HO-1 as aputative protection factor against monocyte apoptosis. Weelucidated the involvement of caspase-3, since caspase-3

Correspondence to: D. Lang; E-mail: langd@uni-muenster.de

Transmitting editor: T. Hunig Received 23 March 2004, accepted 11 November 2004

International Immunology, Vol. 17, No. 2, pp. 155–165 ª 2004 The Japanese Society for Immunologydoi:10.1093/intimm/dxh196 Advance Access publication 20 December 2004

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

performs a number of executive functions leading toapoptosis.

Methods

Antibodies and reagents

The following materials were obtained: phycoerythrin (PE)-conjugated mAb against activated caspase-3 (#557091)(Pharmingen, San Diego, CA); caspase-3 inhibitor (CPP 32,DEVD-CHO) (Calbiochem, Bad Soden, Germany); PE-conjugated mouse anti-human Leu M3 mAb (anti CD14,clone P9, IgG2b) and control mAb of appropriate Ig isotypesand antibodies against bcl-2 (clone bcl-2/100), HO-1 (cloneHO-1-1), CD95 (clone DX2) and CD86 (clone 2331) (BectonDickinson, Palo Alto, CA); FITC-labeled annexin V (BenderMedsystems, Vienna, Austria); FITC-labeled F(ab9)2 frag-ments of goat anti-mouse IgG (Dianova, Hamburg, Germany);caspase-3 antibody (#9662) detecting inactive pro-caspase-3, for western blotting analysis (Cell Signaling, Beverly, MA);reverse transcriptase (Stratagene, Heidelberg, Germany); TaqDNA polymerase (Gibco BRL, Karlsruhe, Germany); dNTPs(New England Biolabs, Beverly, MA); pd(N)6 (BoehringerMannheim, Mannheim, Germany); agarose (AGS, Heidelberg,Germany). Unless otherwise indicated, all other reagents, aswell as hemin (# H-5533), were obtained from Sigma ChemicalCo. (St Louis, MO).

Monocytes and cell culture

Human monocytes were isolated from the leukocyte buffy coatsof healthyvolunteers.Mononuclear cellswereobtainedbyFicoll-Hypaque density gradient centrifugation (400 g, 20 min) fol-lowing which they were washed and then purified further bycentrifugation on a hypotonic Percoll density gradient [57%in phosphate buffered saline (PBS); 400 g, 30 min]. Twointerphases were detected whose upper phase contained en-riched monocytes. Cells were harvested, washed three timesin cold PBS and seeded out into 24-well culture dishes (Greiner,Nurtingen, Germany) in RPMI 1640 culture medium (CM) con-taining 2 mM L-glutamine, 50 lg/ml penicillin/streptomycin,5 mM HEPES and 10 lM mercaptoethanol at 37�C in a 5%CO2/95% air atmosphere. Monocytes were further purifiedby adherence to the culture dishes, resulting in a final purity of>85%. This was assessed by flow cytometry on a FACScan flowcytometer (Becton Dickinson, Mountain View, CA) defined byforward and side light scatter properties, as well as by stainingfor nonspecific esterase and detection of the CD14 surfacemolecule. Monocytes (0.5 3 10/ml) were incubated in CM with5% fetal calf serum (FCS; endotoxin content <0.01 ng/ml) orwere rendered apoptotic by serum starvation with 0.2% FCSfor the indicated culture time. The results were not affected byLPS contamination since all reagents showed <0.001 EU/mlendotoxin concentration on limulus amebocyte lysate assay(BioWhittaker, Walkersville, MD).

Detection of apoptosis

Flow cytometric quantification of apoptosis. Monocyte apo-ptosis was detected either with propidium iodide (PI) nucleistaining or through the use of flow cytometry which detects

CD14 expression and annexin V binding as describedpreviously (12,14). Simultaneous detection of CD14 expres-sion and annexin V binding identifies apoptotic monocytes.We have shown in previous studies that monocytes down-regulate CD14 expression before becoming apoptotic (14).Monocytes were deemed to be apoptotic cells when CD14expression was low and annexin V binding was high afterlymphocytes were gated out flow cytometrically by forwardand side light scatter properties. Monocytes prepared andtreated as described above were double labelled with PE-conjugated Leu M3mAb and annexin V–FITC in staining buffer(SB containing 1% BSA in 50 mM HEPES buffer, pH 7.4) for15 min on ice. PE- and FITC-conjugated murine IgG mAb ofunrelated specifities were used as controls. Cells were washedafter staining and fixed in 4% paraformaldehyde to apply flowcytometry for a total of 10 000 events. Apoptosis was confirmedby DNA electrophoresis and TEM.

Transmission electron microscopy (TEM). Typical morpho-logic changes indicative of apoptosis were evaluated byelectron microscopy as described previously (13). Cells werewashed off the culture dishes, centrifuged, fixed in 1%glutaraldehyde/0.1 M Na cacodylate–HCl, pH 7.4 and post-fixed in 1%OsO4/0.15 M Na cacodylate–HCl, pH 7.4. Sampleswere dehydrated in an ascending ethanol series andembedded in epoxy resin (Epon 812). Ultrathin sectionswere mounted on 150 mesh Formvar-coated copper grids,post-stained with aqueous saturated uranyl acetate and 2%lead citrate, and were then examined using a Philips CM10 electron microscope (Philips Electronics, Mahway, NJ) atan accelerating voltage of 60 kV.

Detection of apoptosis using DNA electrophoresis. DNAextraction and electrophoresis were performed as describedpreviously (14). 1 3 I06 monocytes were lysed by a hypotoniclysing buffer (10 mM Tris, 1 mM EDTA and 0.2% Triton X-100,pH 7.4). After centrifugation (13 000 g, 30 min), supernatantscontaining cleaved chromatin were treated with RNase(50pg/ml)andproteinaseK(100pg/ml).DNAwas thenextractedusing phenolichloroform. After precipitation was performedwith �20�C ethanol and the samples were dried and heated,equal amounts of DNA were placed on a 1% agarose geland separated by electrophoresis for 1 h at 80 V. The lowerdetection limit for visualization of oligonucleosomal bands was1.0 pg of DNA.

Evaluation of cell necrosis

Monocyte viability following all numbers of treatments wasdetermined by trypan blue exclusion or PI uptake of non-permeabilized cells using flow cytometry.

Flow cytometric detection of HO-1, caspase-3, CD95,bcl-2 and CD86

Expression of the aforementioned receptors and proteins wasmeasured by flow cytometry. Monocytes were washed in PBS,fixed with 4% paraformaldehyde and permeabilized with 0.1%saponin for the purposes of detecting caspase-3, bcl-2 andHO-1. For staining, mAbs against activated caspase-3 (PE-conjugated rabbit anti-human), HO-1 [mouse anti-human,

156 Heme oxygenase-1 expression in monocytes

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

unlabeled IgG1, secondary labeling with affinity pure F(ab9)2goat anti-mouse IgG], bcl-2 (FITC-labeled mouse anti-human), CD95 (PE-conjugated mouse anti-human) or CD86(FITC-conjugated mouse anti-human) (5 lg/ml each) orisotype-matched control murine mAb were applied for 20 min.

Indirect immunofluorescence analysis of HO-1 expressionin monocytes

Monocytes were plated on sterile coverslips (LAB-Tek II, Nunc,Wiesbaden) in 24-well plates at 2.5 3 105 cells/ml. After 60 hof culture, adherent monocytes were fixed in 500 ll methanol.Following a further three washes in PBS, the coverslips wereincubated for 10 min in heat-inactivated FCS serum (1/10)to block non-specific antibody binding to Fc receptors. Tovisualize HO-1 in monocytes, cells were then incubated atroom temperature 2 h with mAb specific for HO-1. The cellswere washed and then labeled by incubating for 90 min withaCy3 labeled secondaryantibody.After a further threewashes,cells were incubated with DAPI reagent, washed again andexamined under oil immersion microscopy using a immuno-fluorescence microscope (BX 60, Olympus optical) with a3-CCD camera (DXC-950P, Sony, Koln).

Western blotting analysis

The cells were washed in PBS and resuspended at 10 cells/100 ll of sample buffer containing 2% SDS, 62.5 mM Tris HCl(pH 6.8), 10% glycerol, 5% b-mercaptoethanol and bromphe-nol blue. The cells were then heated at 95�C for 10 min andstored at �20�C until analysis. Each sample contained 20 lgprotein. The samples were separated by a NuPAGE gel 4–12%BT (Invitrogen, Groningen, The Netherlands) and transferredto nitrocellulose membranes [BA83 (0.2 lm) Schleicher undSchull, Dassel, Germany] using a western blotting module(Invitrogen). The membranes were washed three times withpotassium phosphate buffer (0.05 M K3PO4, pH 8.5) and werethen incubated with DIG-NHS-ester (digoxigenin-3-O-methyl-carbonyl-e-aminocapron acid-N-hydroxy-succinimidester) and

Nonidet P 40 (0.01% v/v) for 2 h. After being blocked withcasein blocking reagent (Bio-Rad) for 30 min, the membraneswere incubated with anti-caspase-3 antibody (1:1000). Fol-lowing incubation with anti rabbit IgG peroxidase for 2 h atroom temperature and incubation with Pierce SupersignalFemto West, reactive bands were visualized for the purposesof detecting caspase-3. Mouse IgG2b was used as an isotypematched control for primary antibody. Visualized bands wereanalysed using a Roche Lumi-imager F1.

Inhibition of caspase-3

A specific caspase-3 inhibitor (CPP 32, DEVD-CHO) was usedat a concentration of 100 nM for the purposes of inhibitingcaspase-3. Cells were prepared as described above and in-cubated during the entire culture period.

Semiquantitative RT–PCR

Post-stimulation RNA isolation from monocytes was conductedusing anRNeasyKit (Quiagen,Hilden,Germany) in accordancewith the manufacturer’s instructions. DNase (DNase I, RNase-free,BoehringerMannheim,Germany) digestionwasperformedprior to transcription to cDNA, which was synthesized followingthe addition of 5 lM random primers [pd(N)6, Roche Diagnos-tics,Mannheim,Germany], 1mMdNTPs (NEB,Beverly,MA)andincubation at 37�C with moloney murine leukemia virus reversetranscriptase (Stratagene, Heidelberg, Germany). Contamin-ationwithDNAwasexcludedbyperformingPCR from templatesincubated without reverse transcriptase. The primers used forPCR amplification were 59-ATGGATGATGATATCGCCGCG-39and59-TCTCCATGTCGTCCCAGTTG-39 (humanb-actin, 248bp)(13), as well as 59-AAGATTGCCCAGAAAGCCCTGGAC-39and 59-AACTGTCGCCACCAGAAAGCTGAG-39 (human HO-1,399 bp) (15). The PCR reaction mixture (40 ll) contained 2 mMMgCl2, 0.2mMdNTP,1lMprimerand1UTaqDNApolymerase.Samples were amplified by means of 30 cycles of 60 sdenaturation at 94�C, 30 s annealing at 60�C (HO-1) or 55�C(b-actin) and60selongation in aPeltier thermal cycler (BiometraUno II Thermocycler, Biometra, Gottingen, Germany).The correlation between the expression levels of b-actin and

HO-1 was analysed for semiquantitative PCR. Signal intensityas measured by PCR products was analysed on a 1.5%agarose gel and visualized by ethidium bromide staining.Densitometric quantification of PCR signals was performedusing the Bio Image Intelligent Quantifier program (Bio Image,Ann Arbor, MI).

Statistical analysis

Results are given as means6 SEM. The Mann–Whitney U testand (for paired comparisons) Wilcoxon signed rank test orStudent’s t-test were applied. P < 0.05 was deemed to bestatistically significant. All experiments were performed atleast five times with different buffy coats.

Results

Hemin-induced down-regulation of monocyte apoptosis

Monocyte apoptosis was highly suppressed by heminstimulation (Fig. 1). Apoptosis was quantified by the detection

*

*

*

0

01

02

03

04

05

06

07

08

SCF

h+ nime SPL+ h+ nime

op+ lym nixy B

%5 %2.0 %2.0 %2.0 %2.0

% a

po

pto

tic

c

ells

Fig. 1. Down-regulation of monocyte apoptosis by hemin. The amountof apoptotic monocyte was measured by flow-cytometric staining ofCD14 and annexin V. Under control conditions (5% FCS) 7.06 3.6% ofcells were apoptotic. Serum reduction (0.2% FCS) induced apoptosisrates of 55.7 6 14.5%. This was significantly reduced by stimulationwith hemin (10.9 6 3.6% apoptosis) which yielded survival ratescomparable with those for LPS as a monocyte activator suppressingapoptosis (2.8 6 1.1% apoptosis). LPS contamination was excludedsince polymyxin B did not revert the effects of hemin. These data aremeans 6 SEM from 12 independent experiments (*P < 0.05).

Heme oxygenase-1 expression in monocytes 157

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

of CD14 and annexin V. Cellular apoptosis was induced byserum starvation and detected after 60 h of cell culture. Undercontrol conditions (5% FCS), 7 6 3.6% of monocytes wereapoptotic. Serum reduction to 0.2% FCS yielded 55.76 14.5%apoptotic cells. A hemin concentration of 2.5 lM down-regulated monocytic apoptosis (10.9 6 3.6% apoptoticmonocytes, P < 0.05 compared with 0.2% FCS). LPS (1 lg/ml)was used as a control and reduced apoptotic monocytes to2.8 6 1.1%.Comparable findings were obtained from the evaluation of

monocytes via DNA electrophoresis and TEM (Fig. 2). Heminreduced DNA laddering following serum starvation (Fig. 2A).Figure 2(B) shows representative TEM images of monocytes

under different stimulation conditions. Approximately two-thirds of the monocytes in each visual field were apoptoticfollowing serum starvation. Hemim stimulation reduced apop-totic monocytes to 10–15% of cells per visual field (P < 0.05).Although only negligible numbers of cells showed morpho-logical changes characteristic of apoptosis, some specificchanges following hemin stimulation under serum starvationwere readily apparent. The cells formed tight clumps withsome focus on nucleus staining. However, this did not impededetection of surface molecules, since cells were smoothlydetached from each other for FACS analysis. The cells showedthe same CD14 expression levels as for the control cells (5%FCS). Thus, it is possible that morphological changes are not

Fig. 2. Down-regulation of apoptosis after stimulation with hemin detected by DNA electrophoresis and TEM. (A) Oligonuclosomal DNA cleavageof monocytes after hemin (2.5 lM) treatment. Electrophoresis of DNA isolated from 1 3 106 monocytes was performed after 48 h culture. Formedium control (5% FCS), no nucleosomal DNA fragments were detected after ethidium bromide staining. After serum starvation (0.2% FCS),typical oligonucleosomal DNA laddering was found. Oligonucleosomal DNA fragments were significantly reduced after hemin (2.5 lM) treatment.(B) TEM demonstrates protection from apoptosis by hemin. The left section depicts twomonocytes that maintained normal morphology after a 60 hculture in 5% FCS containing medium. The middle panel shows typical changes of two apoptotic monocytes after serum deprivation for 60 h.Dense nuclear condensation, cytoplasmic vacuolization, shrinkage and rounding are visible in the upper monocyte. The lower monocyte shows anearly stage of apoptosis with incipient nuclear condensation. Less than 15% of the monocytes analysed showed typical apoptotic features pervisible field after serum deprivation and concurrent stimulation with hemin (2.5 lM; right panel). No typical changes of apoptosis are discernible,although some unspecific changes are remarkable. The cells clumped and had had some emphasis in nuclear staining. These were unspecificchanges owing to stress reaction. Magnification is 37200.

158 Heme oxygenase-1 expression in monocytes

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

attributable to apoptosis and that they are instead elicited byunspecificstress-inducedchangessecondary toserumstarva-tion and hemin stimulation. All experiments were performed atleast five times.

Time- and dose-dependent induction of monocytic HO-1by hemin at mRNA and protein levels

Figure 3(A and B) shows that HO-1 mRNA was induced byhemin in monocytes in a dose- (A) and time-dependent (B)manner. This was done by semiquantitative RT–PCR wherebythe resulting means 6 SEM of arbitrary units after densito-metric quantification of PCR signals are shown in a graph ofone representative RT–PCR blotting for five experiments(Fig. 3A). Signal intensity was compared to the b-actin levelanalysed by semiquantitative RT–PCR. Hemin induced HO-1significantly from the basal level of 1 arbitrary unit after 24 h ofincubation time up to 2.7 6 1.6 units (1 lM hemin; P < 0.05).HO-1 mRNA increased continuously with increased dosagesof hemin and led to a maximum mRNA intensity of 9.4 6 3.4units with 10 lM hemin.Since 2.5 lM hemin increased HO-1 mRNA significantly, this

concentration was used to investigate time dependency. HO-1mRNA peaked at 24 h of incubation and declined after 48 h.After 2 h, only a moderate increase in HO-1 mRNA (1.3 6 0.1arbitrary units) was detected. A significant increase wasobserved after 12 h, resulting in 4.7 6 2.0 arbitrary units, andthis increase nearly doubled after 24 h (7.9 6 2.2 arbitraryunits) (*P < 0.05; Fig. 3B).HO-1 regulation at protein level was comparable to mRNA

expression (Fig. 3C and D). One-micromolar hemin stimulatedHO-1 protein expression significantly from a basal level of11.9 6 5.7% cells expressing HO-1 up to 40.2 6 13.3%, asquantified by FACS analysis (P < 0.05). Time dependency wasdetected at protein level at different times than mRNA, sinceprotein expression has to be expected during a later timecourse (E and F). Flow-cytometric analysis revealed that HO-1protein levels increased continuously with increased dosagesof hemin and peaked after 16 h of hemin stimulation (2.5 lM)(70 6 9% HO-1 positive cells) (*P < 0.05; Fig. 3C and E).

Hemin-induced caspase-3 and the effects on apoptosis ofadditional caspase-3 inhibition

Our analysis of caspase-3 expression as one of the maineffectors of apoptosis revealed that hemin significantly in-duced activated caspase-3. Culturing cells in 0.2% FCScontaining medium caspase-3 positive cells increased from236 4% (0.2% FCS) to 56.66 6% (2.5 lM, P < 0.05) (Fig. 4A).In order to rule out the possibility that the increase in theamount of activated caspase-3 positive cells was elicited byincreased precursor induction, we performed western blottinganalysis against pro-caspase-3 (Fig. 4C). Pro-caspase-3 wasdown-regulated following hemin stimulation, thus demonstrat-ing that the increased caspase-3 level was mainly caused byactivated/active caspase-3. The fact that caspase-3 wasblocked by the specific inhibitor DEVD-CHO showed thathemin-induced suppression of the apoptosis rate could befurther down-regulated (P < 0.05; Fig. 4D). Caspase-3inhibition alone did not affect apoptosis induction followingserum starvation with the concentration applied (56 6 15%

Fig. 3. Dose and time dependency of HO-1 mRNA expression (A andB) and of HO-1 protein expression (C–G) after hemin treatment. HO-1is induced by hemin in monocytes in a dose- (A) and time-dependentfashion (B) at the mRNA and protein level (C–G). Hemin at aconcentration of 1 lM induced HO-1 significantly after 24 h of in-cubation time (P < 0.05). HO-1 mRNA increased continuously withincreased dosages of hemin, resulting in a maximum of mRNAintensity of 9.4 6 3.4 units with 10 lM hemin (A). HO-1 mRNA peakedat 24 h of incubation and decreased again after 48 h. A significantincrease was measurable after 12 h, resulting in 4.7 6 2.0 arbitraryunits. This enhancement in expression nearly doubled after 24 h (7.962.2 arbitrary units) (*P< 0.05) (B). HO-1 signal intensity was comparedto b-actin analysed by semiquantitative RT–PCR. One representativeRT–PCR blotting is shown out of five experiments. The HO-1 proteinlevel showed comparable regulation to mRNA (C–F). Hemin (1 lM)stimulated HO-1 protein expression significantly detected by FACSanalysis. Time dependency was detected at protein level at differenttimes than mRNA since protein expression has to be expected duringa later time course. (C and D) Quantitation and statistical analysis ofeight experiments are presented in (F) for different hemin concen-trations and in (E) for time dependancy (n ¼ 5). For each of them onerepresentative FACS analysis is shown in (D) and (F), respectively. HO-1 protein expression peaked at 16 h of stimulation with hemin (*P <0.05). Induction of HO-1 was also detected by indirect immunofluo-rescence after 60 h (G). Without stimulation nearly no HO-1 expressionwas detectable (red signals, upper panel). Stimulation with 1 lMhemin(middle) and 2.5 lM hemin (lower panel) resulted in increasing signalsfor HO-1 expression and confirmed FACS Data. Iso ¼ isotype control.

Heme oxygenase-1 expression in monocytes 159

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

apoptosis without DEVD-CHO as compared to 51 6 18% withDEVD-CHO). However, this concentration was combined witha hemin treatment that has the capacity to reduce apoptosis.

Effect of the HO-1 antagonist zinc protoporphyrine (ZnPP)on apoptosis

HO-1 inhibition by ZnPP, a competitive blocker of HO-1, led toa 306 9% increase in the apoptosis rate in monocytes treatedwith hemin (P < 0.05) whereas ZnPP reversed the anti-apoptotic effects of HO-1, with 10.9 6 3.6% of the cellsbecoming apoptotic following serum deprivation and HO-1induction via hemin alone (Fig. 5). Dexamethasone-inducedapoptosis showed apoptosis rates of 83.5 6 2% that werenot significantly altered by ZnPP. Hemin significantly reduced

apoptosis due to the addition of dexamethasone (25 6 5.7%apoptosis) and ZnPP again reversed this effect, leading tohigh levels of apoptosis (78.5 6 3.5% apoptosis, P < 0.05).These results once again show that ZnPP can block hemin-induced anti-apoptotic pathways, and this in turn points to thecytoprotective effects of HO-1.

Detection of CD95, bcl-2 and CD86

The primary CD95 death receptor is significantly down-regulated by hemin (22 6 4% CD95 positive monocyteswithout hemin versus 106 3% CD95 positive cells with hemin;P < 0.05) (Table 1). No significant differences were found forbcl-2, which is an important member of a family of proteins thatprevents apoptosis. On the other hand, the costimulatory

Fig. 3. Continued.

160 Heme oxygenase-1 expression in monocytes

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

receptor CD86 significantly increased following hemin stimu-lation (4 6 4% CD86 positive cells without hemin versus 19 6

3% CD86 positive cells with hemin; P < 0.05), thus indicatingthat hemin is associated with the induction of monocyte dif-ferentiation into macrophages. Furthermore, we could showthat the inhibition of HO-1 by ZnPP reverted the induction ofCD86 by hemin. The same was true for inhibition of caspase-3which also inhibited the expression of hemin-induced cos-timulatory CD86 (Fig. 6).

Discussion

Monocytes/macrophages are centrally involved in the patho-genesis of various inflammatory and vascular disorders suchas sepsis, atherosclerosis and transplant rejection (16–18).Recent data from organ transplantation, xenotransplantation,ischemia/reperfusion injury and inflammatory kidney disordersreveals that the HO-1 enzyme diminishes inflammation andimproves outcome in these disorders by mediating cytopro-tection in various cell types (19–21). Most studies have in-vestigated the effects of HO-1 on cardiac myocytes,

endothelial cells or smooth muscle cells (22–24). Few dataare available on monocytes (25). HO-1 has become a focalpoint in a number of animal studies because it is easilyinducible by its own substrate hemin/heme. HO-1 can also beblocked by metallo-protoporphyrins, all of which have minortoxic effects when pharmacological doses are applied in vivo(26–29). In view of the crucial role played by monocytes,particularly in the aforementioned settings, we elucidated theeffects of HO-1 in freshly cultured cells on the mechanismsand regulation of apoptosis.As was found for other cell types (2), hemin is a highly

effective HO-1 inducer that antagonized PCD effectively inmonocytes rendered apoptotic by serum deprivation ordexamethasone (Figs 1 and 5). We found that this cytopro-tective effect is reversed by the specific competitive HO-1inhibitor, ZnPP. This indicates that hemin-enhanced life span isassociated with induction of active HO-1, establishing a directlink between HO-1 up-regulation and cytoprotective effects.Hemin up-regulated monocyte HO-1 dose-dependently on

the mRNA and protein levels. HO-1 peaks were obtained after16 and 24 h, but apoptosis levels could only be measuredafter 60 h owing to the fact that the cells did not induceapoptotic changes before that time. We found that the anti-apoptotic effect of hemin treatment was most pronouncedwhen hemin was added to the culture medium within 24 h.Hemin added at a later time was less protective (data notshown). We also tried to ascertain which HO-1-dependentmetabolic products (iron, CO and bilirubin or biliverdin) triggercytoprotection in monocytes. Since iron sulfate/ferric gluco-nate and the iron chelator desferrioxamine had no effect onhemin-induced protection from PCD, iron-related effects couldbe ruled out. CO and bilirubin were not measurable in ourculture supernatants or cell lysates due to the extremely lowconcentrations or short half-life times of these substances.Thus, we were unable to elucidate the role of these twometabolites. However, exogenously administered bilirubinsignificantly inhibited apoptosis of monocytes (data notshown), which suggests that this substance may mediatehemin-dependent cytoprotection. Other key mechanismssuch as reactive oxygen intermediate pathways may alsoplay a role in the regulation of spontaneous monocyte apop-tosis blocked by HO-1. These pathways require furtherinvestigation. Surprisingly, our additional investigations ofhemin-induced changes revealed that, despite apoptosisinhibition, caspase-3 expression was up-regulated in mono-cytes. Caspase-3 activation by HO-1 has recently beendemonstrated for vascular smooth muscle cells, but here,adenovirus-mediated HO-1 stimulation triggered apoptosisinstead of abrogating it (30). Hemin activated caspase-3 ef-fectively in monocytes, which was proven by co-administrationof the specific caspase-3 antagonist DEVD-CHO which furtherenhanced cell protection and the life span of monocytes inour investigation. In our view, the effects of hemin on apoptosisare regulated in a cell-specific manner and HO-1 may not havea protective effect in all types of cells.Our findings demonstrate that the effects of hemin are not

limited to HO-1 induction and caspase-3 activation; in ourstudy, apoptosis-mediating receptors such as CD95 were alsodown-regulated (Table 1). This inhibition could be a principalcytoprotective effector in settings involving the CD95/CD95L

Fig. 3. Continued.

Heme oxygenase-1 expression in monocytes 161

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

Fig. 4. Induction of activated caspase-3 in monocytes by hemin (A and B) was paralleled by down-regulation of inactive pro-caspase-3 (C) whencaspase-3 antagonist DEVD-CHO influenced hemin-dependent monocyte apoptosis (D). (A and B) Hemin significantly induced activatedcaspase-3. At 0.2% FCS containing medium, 23 6 4% caspase-3 positive monocytes increased to 56.7 6 6% by hemin stimulation (2.5 lM, P <0.05). Monocytes cultured in 5% FCS served as control. A maximum increase was noted for 5 lM hemin. (B) Representative FACS data are shownfor which quantification and statistical analyses are presented in (A). (C) Proving results generated by FACS analysis, western blotting in order todetect pro-caspase-3 with a specific antibody against inactive, uncleaved pro-caspase-3 was performed. Hemin down-regulated pro-caspase-3after serum starvation (0.2% FCS and hemin) which demonstrated that FACS analysis detected activated caspase-3 and not inactive pro-caspase.(D) Blocking of active caspase-3 by the specific inhibitor DEVD-CHO (100 nM) showed that inhibition of apoptosis after hemin treatment could befurther down-regulated (*P < 0.05, B). The relevant level of active caspase-3 was detected after 16 h, apoptosis after 60 h.

162 Heme oxygenase-1 expression in monocytes

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

system, as has been shown for glucocorticoid-induced mono-cyte PCD (31). Since hemin effectively counter-regulatesapoptosis following serum starvation—a process that doesnot involve CD95/CD95L to any significant degree—otherreceptors and pathways must be involved. In our supplemen-tary morphological studies of hemin-treated monocytescultured for more than 60 h, rapid differentiation into macro-phages was observed (Table 1), nuclei shape altered,cytoplasmic volume increased (Fig. 2) and expression of thecostimulatory receptor CD86 (which is characteristic ofmacrophage activation) rose significantly (Table 1). Classicdendritic cell markers such as CD1a remained unaffected.This monocyte to macrophage differentiation might be attribut-able to hemin-induced caspase-3 up-regulation and HO-1induction since specific inhibition of caspase-3 and HO-1,respectively, blocked the induction of CD 86 (Fig. 6).

A recent report revealed that the principal monocyte growthfactor M-CSF induces differentiation of monocytes via cas-pase activation and apoptosis suppression involving cyto-chrome c release from mitochondria (32). In contrast toM-CSF-induced differentiation, hemin-dependent differenti-ation might be promoted by iron, which is released by HO-1activation and has been attributed to monocyte into macro-phage differentiation (33,34). Our findings are of particular

Fig. 5. Influence of ZnPP as an HO-1 blocker on monocyte apoptosisinduced by serum starvation alone or with dexamethasone treatment.ZnPP as a competitive blocker of HO-1 reversed the anti-apoptoticeffects of HO-1 induced by hemin. Apoptosis increased significantlyfrom 7.0 6 3.0% to 32 6 10% by ZnPP (10 lM) in 5% FCS-containingmedium. The amount of apoptosis was not affected by ZnPP afterserum starvation. A significant reduction of apoptosis was found inmonocytes after hemin treatment, an effect which was significantlyreverted again by ZnPP. Very high apoptosis levels achieved by 0.2%FCS medium plus dexamethasone could be highly suppressed byhemin and also reverted again by ZnPP. Dexa: dexamethasone ata concentration of 10 lM. The hemin concentration used was 2.5 lM.ZnPP: zinc protoporphyrine (10 lM). (*P < 0.05 compared to stimu-lation without ZnPP, + P < 0.05 compared to 0.2% FCS alone.)

Table 1. Regulation of CD95, bcl-2 and CD86 by hemintreatment

CD95 positivecells (%)

bcl-2 positivecells (%)

CD86positivecells (%)

Control (5% FCS) 23 6 6 14 6 6 11 6 20.2% FCS 22 6 4 9 6 4 4 6 40.2% FCS +hemin (2.5 lM)

10 6 3* 14 6 7 19 6 3*

Monocytes were treated with hemin (2.5 lM) for 60 h for the determin-ation of CD95, and bcl-2 and for 168 h for the determination of CD86.Data are means 6 SEM, n ¼ 5; *P < 0.05 as compared to 0.2% FCSalone. Fig. 6. Induction of CD86 in monocytes by hemin was abrogated by

ZnPP as an HO-1 blocker and the specific caspase-3-inhibitor DEVD-CHO. Amarker for differentiation in humanmonocytes is CD 86. Hemincould induce significantly the expression of CD 86 after 168 h ofincubation whereas ZnPP (10 lM) as an HO-1 blocker and the specificinhibitor Caspase-3-Inhibitor (C-3-Inhib) DEVD-CHO (100 nM) them-selves did not change expression of CD 86 compared to unstimulatedmonocytes (ctrl ¼ control). ZnPP, as well as the caspase-3-inhibitor,could also prevent the expression of CD86 after stimulation by hemin.Influence of relevant endotoxin contamination was excluded by add-itional polymyxin D stimulation. (A) Four experiments are summarized.(B) A representative histogram.

Heme oxygenase-1 expression in monocytes 163

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

significance in light of recent reports on the beneficial effectsof HO-1-dependent anti-inflammation or tissue protection(19,35,36). In settings where vascular occlusive disorders ororgan transplant rejection are amenable to amelioration byHO-1 up-regulation via smooth muscle cell apoptosis withconcurrent consecutive vascular stenosis desobliteration andendothelial cell protection that promotes re-endothelialization,this improvement could be antagonized by HO-1-specificeffects on monocytes. On the other hand, recent data showthat the influx of monocytes/macrophages are not necessarilysynonymous with tissue damage and progressive inflamma-tion (37). Macrophage function may be more important thanthe number of infiltrating cells. Additional studies need toelucidate the changes in monocyte function by different treat-ments which will also help to understand the regulation of localinflammatory responses.Nevertheless, in atherosclerosis, monocytes are supposed

to aggravate lesions and promote stenosis and vascularremodeling. In organ transplantation, these cells seem toprovoke immunosuppressive-resistant rejection or fibroticinjury (28,38). Supportingmonocyte survival and differentiationinto macrophages, hemin/HO-1 treatment may have more thanjust beneficial effects achieved by effects on intrinsic vascularcells during these processes. Thus, further studies on HO-1-dependent tissue protection should avoid inducing anti-apoptotic effects inmonocytes whereby especially the functionof monocytes has to be taken into account, probablymore thanthe absolute number of infiltrating monocytes.In conclusion, our study shows that hemin can generate

HO-1 in cultured monocytes. Although HO-1 induction antag-onizes apoptosis, there is an increased caspase-3 activity anda distinct induction of cell differentiation into macrophages.Experiments that treat various inflammatory, vascular orimmune diseases with HO-1 should precisely characterizethe effects of this enzyme on all cellular compartmentsinvolved.

Acknowledgements

The authors thank Ms. K. Beul for excellent technical assistance. Thiswork was supported by I2KF Muenster, Project No. D20.

Abbreviations

CM culture mediumFCS fetal calf serumHO heme oxygenase(s)HO-1 heme oxygenase-1PCD programmed cell deathPI propidium iodidePS phosphatidylserineSB staining buffer

References

1 Agarwal, A. and Nick, H. S. 2000. Renal response to tissue injury:lessons from heme oxygenase-1 gene ablation and expression.J. Am. Soc. Nephrol. 11:965.

2 Inguaggiato, P., Gonzalez-Michaca, L., Croatt, A. J., Haggard,J. J., Alam, J. and Nath, K. A. 2001. Cellular overexpression ofheme oxygenase-1 up-regulates p21 and confers resistance toapoptosis. Kidney Int. 60:2181.

3 Dong, Z., Lavrovsky, Y., Venkatachalam, M. A. and Roy, A. K. 2000.Heme oxygenase-1 in tissue pathology. The yin and yang. Am. J.Pathol. 156:1488.

4 Datta, P. K., Koukouritaki, S. B., Hopp, K. A. and Lianos, E. A. 1999.Heme oxygenase-1 induction attenuates inducible nitric oxidesynthase expression and proteinuria in glomerulonephritis. J. Am.Soc. Nephrol. 10:2540.

5 Kacimi, R., Chentoufi, J., Honbo, N., Long, C. S. and Karliner, J. S.2000. Hypoxia differentially regulates stress proteins in culturedcardiomyocytes. Role of the p38 stress-activated kinase signalingcascade, and relation to cytoprotection. Cardiovasc. Res. 46:139.

6 Otterbein, L. E., Kolls, J. K., Mantell, L. L., Cook, J. L., Alam, J.and Choi, A. M. K. 1999. Exogenous administration of hemeoxygenase-1 by gene transfer provides protection againsthyperoxia-induced lung injury. J. Clin. Invest. 103:1047.

7 Akagi, R., Takahashi, T. and Sassa, S. 2002. Fundamental role ofheme oxygenase in the protection against ischemic acute renalfailure. Jpn J. Pharmacol. 88:127.

8 Willis, D., Moore, A. R., Frederick, R. and Willoughby, D. A. 1996.Heme oxygenase: a novel target for the modulation of the inflam-matory response. Nat. Med. 2:87.

9 Nath, K. A., Haggard, J. J., Croatt, A. J., Grande, J. P., Poss, K. D.and Alam, J. 2000. The indispensability of heme oxygenase-1 inprotecting against acute heme protein-induced toxicity in vivo. Am.J. Pathol. 156:1527.

10 Zhang, X., Shan, P., Otterbein, L. E., Alam, J., Flavell, R. A., Davis,R. J., Choi, A. M. K. and Lee, P. J. 2003. Carbon monoxideinhibition of apoptosis during ischemia-reperfusion lung injury isdependent on the p38 mitogen-activated protein kinase pathwayand involves caspase 3. J. Biol. Chem. 278:1248.

11 Stocker, R., Yamamoto, Y., McDonagh, A. F., Glazer, A. N. andAmes, B. N. 1987. Bilirubin is an antioxidant of possible physio-logical importance. Science 235:1043.

12 Lang, D., Hubrich, A., Dohle, F., Terstesse, M., Saleh, H., Schmidt,M., Pauels, H. G. and Heidenreich, S. 2000. Differential expressionof heat shock protein 70 (hsp70) in human monocytes renderedapoptotic by IL-4 or serum deprivation. J. Leukoc. Biol. 68:729.

13 Lang, D., Dohle, F., Terstesse, M., Bangen, P., August, C., Pauels,H. G. and Heidenreich, S. 2002. Down-regulation of monocyteapoptosis by phagocytosis of platelets: involvement of a caspase-9,caspase-3, and heat shock protein 70-dependent pathway.J. Immunol. 168:6152.

14 Heidenreich, S., Schmidt, M., August, C., Cullen, P., Rademaekers,A. and Pauels, H. G. 1997. Regulation of human monocyteapoptosis by the CD14 molecule. J. Immunol. 159:3178.

15 Barber, A., Robson, S. C. and Lyall, F. 1999. Hemoxygenase andnitric oxide synthase do not maintain human uterine quiescenceduring pregnancy. Am. J. Pathol. 155:831.

16 Vayssier-Taussat, M., Camilli, T., Aron, Y., Meplan, C., Hainaut, P.,Polla, B. S. and Weksler, B. 2001. Effects of tobacco smoke andbenzo(a)pyrene on human endothelial cell and monocyte stressresponses. Am. J. Physiol. Heart Circ. Physiol. 280:H1293.

17 Lessner, S. M., Prado, H. L., Waller, E. K. and Galis, Z. S. 2002.Atherosclerotic lesions grow through recruitment and proliferationof circulating monocytes in a murine model. Am. J. Pathol. 160:2145.

18 Osterud, A. and Bjorklid, E. 2003. Role of monocytes inatherogenesis. Physiol. Rev. 83:1069.

19 Tullius, S. G. et al. 2002. Inhibition of ischemia/reperfusion injuryand chronic graft deterioration by a single-donor treatment withcobalt-protoporphyrin for the induction of heme oxygenase-1.Transplantation 74:591.

20 Soares, M. P., Brouard, S., Smith, R. N. and Bach, F. H. 2001. Hemeoxygenase-1, a protective gene that prevents the rejection oftransplanted organs. Immunol. Rev. 84:275.

21 Katori, M., Busuttil, R. W. and Kupiec-Weglinski, J. W. 2002. Hemeoxygenase-1 system in organ transplantation. Transplantation 74:905.

22 Tulis, D. A., Durante,W., Peyton, K. J., Evans, A. J. and Schafer, A. I.2001. Heme oxygenase-1 attenuates vascular remodelingfollowing balloon injury in rat carotid arteries. Atherosclerosis155:113.

164 Heme oxygenase-1 expression in monocytes

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from

23 Duckers, H. J. et al. 2001. Heme oxygenase-1 protects againstvascular constriction and proliferation. Nat. Med. 7:693.

24 Liu, X. M., Chapman, G. B., Peyton, K. J., Schafer, A. I. andDurante, W. 2002. Carbon monoxide inhibits apoptosis in vascularsmooth muscle cells. Cardiovasc. Res. 55:396.

25 Immenschuh, S., Stritzke, J., Iwahara, S. and Ramadori, G. 1999.Up-regulation of heme-binding protein 23 (HBP23) gene ex-pression by lipopolysaccharide is mediated via a nitricoxide-dependent signalingpathway in ratKupffer cells.Hepatology30:118.

26 Weber, T. P., Meissner, A., Boknik, P., Hartlage, M. G., Mollhoff, T.,Van Aken, H. and Rolf, N. 2001. Hemin, inducer of heme-oxygenase 1, improves functional recovery from myocardialstunning in conscious dogs. J. Cardiothorac. Vasc. Anesth. 15:422.

27 Andersson, J. A., Uddman, R. and Cardell, L. O. 2002. Hemin,a heme oxygenase substrate analog, both inhibits and en-hances neutrophil random migration and chemotaxis. Allergy 57:1008.

28 Ishikawa, K., Navab, M., Leitinger, N., Fogelman, A. M. and Lusis,A. J. 1997. Induction of heme oxygenase-1 inhibits the monocytetransmigration induced by mildly oxidized LDL. J. Clin. Invest.100:1209.

29 Yang, G., Nguyen, X., Ou, J., Rekulapelli, P., Stevenson, D. K. andDennery, P. A. 2001. Unique effects of zinc protoporphyrin on HO-1induction and apoptosis. Blood 97:1306.

30 Liu, X. M., Chapman, G. B., Wang, H. and Durante, W. 2002.Adenovirus-mediated heme oxygenase-1 gene expression stimu-lates apoptosis in vascular smooth muscle cells. Circulation105:79.

31 Schmidt, M., Lugering, N., Lugering, A., Pauels, H. G., Schulze-Osthoff, K., Domschke, W. and Kucharzik, T. 2001. Role of theCD95/CD95 ligand system in glucocorticoid-induced monocyteapoptosis. J. Immunol. 166:1344.

32 Sordet, O., Rebe, C., Plenchette, S., Zermati, Y., Hermine, O.,Vainchenker, W., Garrido, C., Solary, E. and Dubrez-Daloz, L. 2002.Specific involvement of caspases in the differentiation of mono-cytes into macrophages. Blood 100:4446.

33 Gazitt, Y., Reddy, S. V., Alcantara, O., Yang, J. and Boldt, D. H.2001. A new molecular role for iron in regulation of cell cycling anddifferentiation of HL-60 human leukemia cells: iron is required fortranscription of p21(WAF1/CIP1) in cells induced by phorbolmyristate acetate. J. Cell Physiol. 8:124.

34 Asada, M., Yamada, T., Ichijo, H., Delia, D., Miyazono, K.,Fukumuro, K. and Mizutani, S. 1999. Apoptosis inhibitory activityof cytoplasmic p21(Cip1/WAF1) in monocytic differentiation.EMBO J. 18:1223.

35 Oh, G. S., Pae, H. O., Moon, M. K., Choi, B. M., Yun, Y. G., Rim, J. S.and Chung, H. T. 2003. Pentoxifylline protects L929 fibroblastsfrom TNF-alpha toxicity via the induction of heme oxygenase-1.Biochem. Biophys. Res. Commun. 302:109.

36 Datta, P. K., Gross, E. J. and Lianos, E. A. 2002. Interactionsbetween inducible nitric oxide synthase and heme oxygenase-1 inglomerulonephritis. Kidney Int. 61:847.

37 Kluth, D. C., Erwing, L.-P. and Rees, A. J. 2004. Multiple facets ofmacrophages in renal injury. Kidney Int. 66:542.

38 Heidenreich, S., Lang, D., Tepel, M. and Rahn, K. H. 1994.Monocyte activation for enhanced tumour necrosis factor-alphaand interleukin 6 production during chronic renal allograftrejection. Transpl. Immunol. 2:35.

Heme oxygenase-1 expression in monocytes 165

by guest on Decem

ber 22, 2014http://intim

m.oxfordjournals.org/

Dow

nloaded from