Compartmentalization of lipid peroxidation in sepsis by multidrug-resistant gram-negative bacteria: experimental and clinical evidence.

0.67 MB·English

by  Toufekoula, Chryssoula

Toufekoulaetal.CriticalCare2013,17:R6
http://ccforum.com/content/17/1/R6
RESEARCH Open Access
Compartmentalization of lipid peroxidation in
sepsis by multidrug-resistant gram-negative
bacteria: experimental and clinical evidence
Chryssoula Toufekoula1, Vassileios Papadakis1, Thomas Tsaganos1, Christina Routsi2, Stylianos E Orfanos3,
Anastasia Kotanidou2, Dionyssia-Pinelopi Carrer1, Maria Raftogiannis1, Fotini Baziaka1 and
Evangelos J Giamarellos-Bourboulis1,4*
Abstract
Introduction: Recent evidence suggests a link between excess lipid peroxidation and specific organ failures in
sepsis. No study has been performed in sepsis by multidrug-resistant (MDR) Gram-negative bacteria.
Methods: Lethal sepsis was induced in rats by the intraperitoneal injection of one MDR isolate of Pseudomonas
aeruginosa. Produced malondialdehyde (MDA) was measured in tissues 5 hours after bacterial challenge with the
thiobarbiturate assay followed by high-performance liquid chromatography (HPLC) analysis. Results were compared
with those from a cohort of patients with ventilator-associated pneumonia (VAP) and sepsis by MDR Gram-
negative bacteria. More precisely, serum MDA was measured on 7 consecutive days, and it was correlated with
clinical characteristics.
Results: MDA of septic rats was greater in the liver, spleen, and aortic wall, and it was lower in the right kidney
compared with sham operated-on animals. Findings were confirmed by the studied cohort. Circulating MDA was
greater in patients with hepatic dysfunction and acute respiratory distress syndrome (ARDS) compared with
patients without any organ failures. The opposite was found for patients with acute renal dysfunction. No
differences were found between patients with ARDS without or with cardiovascular (CV) failure and patients
without any organ failure. Serial measurements of MDA in serum of patients indicated that levels of MDA were
greater in survivors of hepatic dysfunction and ARDS and lower in survivors of acute renal dysfunction.
Conclusions: Animal findings and results of human sepsis are complementary, and they suggest a
compartmentalization of lipid peroxidation in systemic infections by MDR gram-negative bacteria.
Introduction elicitedbyoxidativestressoccurringinbiologicmaterial.
Oxidative stress results from an imbalance between pro- Malondialdehyde(MDA)isthemost abundant lipidper-
ductionofreactiveoxygenandnitrogenspecies(ROSand oxidation-specificaldehyde[2].
RNS) and endogenous antioxidant defense mechanisms The process of lipid peroxidation accompanied by
[1].Agrowingbodyofevidencesuggeststhatmanyofthe excess production of MDA has been well documented
effects of cellular dysfunction under oxidative stress are bothinexperimentalstudieswithcellsinfectedbyspeci-
mediatedbyproductsofnonenzymaticreactions,suchas fic pathogens [3]andinsystemic infectionscomplicated
the peroxidative degradation of polyunsaturated fatty by sepsis and leading to death [4,5]. Available evidence
acids.Aldehydemoleculesgeneratedduringlipidperoxi- fromstudiesinhumansevokesthe impressionthatlipid
dationare considered ultimate mediatorsof toxic effects peroxidationisaharmfuleventubiquitouslyaccompany-
ingallcasesofsystemicinflammationandsepsistosuch
anextentthatcirculatingMDAmaybeconceivedofasa
*Correspondence:[email protected]
biomarkerofunfavorableprognosis[4-6].Surprisingly,a
14thDepartmentofInternalMedicine,UniversityofAthens,MedicalSchool,
12462Athens,Greece recent studyinpatientswithseveresepsis(that is,sepsis
Fulllistofauthorinformationisavailableattheendofthearticle
©2013Toufekoulaetal.;licenseeBioMedCentralLtd.ThisisanopenaccessarticledistributedunderthetermsoftheCreative
CommonsAttributionLicense(http://creativecommons.org/licenses/by/2.0),whichpermitsunrestricteduse,distribution,and
reproductioninanymedium,providedtheoriginalworkisproperlycited.
Toufekoulaetal.CriticalCare2013,17:R6 Page2of11
http://ccforum.com/content/17/1/R6
complicated by organ failure) showed that MDA is Pseudomonas aeruginosa, and then we tried to confirm
increasedin the circulation of patientswithhepatic fail- findings of circulating MDA in patients with specific
ure and with renal failure but not in the circulation of organ failures after ventilator-associated pneumonia
patients with respiratory failure and circulatory failure (VAP)fromMDRgram-negativebacteria(Figure1).
[7].Theseresultsweregeneratedinacohortof50sepsis
patients and required further confirmation. However, Materials and methods
they suggest that lipid peroxidation is not a ubiquitous Animal study
phenomenon, and that it may be implicated in specific Intotal,60 male Wistar rats werestudied. Their mean ±
organfailuresinacompartmentalizedfashion. SDweightwas253.4±22.4g.Thestudyreceivedpermis-
Manycasesofsepsisdevelopascomplicationsofhospi- sionfrom theVeterinaryDirectorate of thePerfectureof
talizationinthe intensive care unit(ICU).In suchcases, Athens,accordingtotheGreeklegislation,inconformance
the offending pathogens are often multidrug resistant totheCouncilDirectiveoftheEuropeanCommunity.Rats
(MDR),andtheyimposeconsiderabledifficultiesinman- were housed in metal cages and had access to tap water
agement because the available antimicrobial agents are andstandardbalancedchowadlibitum.Temperatureran-
limited [8]. The only solution to this emerging problem ged between 18°C and 22°C; relative humidity, between
isthein-depthunderstandingofthepathogenesisbehind 55%and65%;andthelight/darkcyclewas6AMto6PM.
sepsis caused by MDR bacteria. At this point, it should One multidrug-resistant (MDR) isolate of P. aeruginosa
be underscored that the studied models of sepsis use fromthebloodofapatientwithseveresepsiswasapplied.
either purified PAMPs (pathogen-associated molecular Minimalinhibitoryconcentrations(MICs)ofpiperacillin/
patterns) or peritonitis after CLP (cecal ligation and tazobactam, ceftazidime, imipenem, meropenem, cipro-
puncture)asastimulus.MDRbacteriaareconsideredto floxacin, and amikacin, determined with a microdilution
stimulate host responses in a fashion similar to that of technique, were 256/4,256,128, 64,512,and512 μg/ml,
community-acquired susceptible bacteria. However, respectively.Theisolatewasstoredasmultiplealiquotsin
muchambiguityexistswhetherMDRgram-negativeiso- skimmilk(OxoidLtd,London,UK)under-70°C.Oneali-
latespossessdifferentvirulencecharacteristicsfromsus- quot was removed from the refrigerator before each
ceptible gram-negative isolates. This is very difficult to experiment.Singlecolonieswereincubatedat37°Cin10
conclude from a clinical study in which inappropriate ml of Mueller-Hinton broth (Oxoid Ltd) for 8 hours to
antimicrobial therapy is the driver for unfavorable out- yielda log-phase inoculumthat was appliedforbacterial
comes after infection with MDR bacteria and does not challenge.
allow focus on precise virulence characteristics [9]. In a Animals were divided into two groups for treatment,
recent retrospective study, the physical course of 29 as follows:
patients infected with carbapenem-resistant Enterobac- (cid:129) Group A (n = 15), sham-operated; injected intraperi-
teriaceae was compared with that of 29 well-matched toneally with 1 ml of Mueller-Hinton; and
patients infected with carbapenem-susceptible Entero- (cid:129) Group B (n = 45), animals injected intraperitoneally
bacteriaceae;in-hospitalmortalitieswere30.8%and7.4%, with 1 ml of the prepared inoculum of the test isolate
respectively (P = 0.04) [10]. Similar results were drawn yielding a challenge of 5 × 106 cfu/rat.
fromaretrospectivestudyofourgroupontheoutcomes Survivalwasrecordedat12-hourintervalsforfourani-
of243patients;howeverwhenpatientswithseveresepsis malsofgroupAandfor10animalsofgroupB. Toiden-
were analysed separately, it wasfoundthat severe sepsis tify the peak time point of lipid peroxidation, animals of
causedbyMDRbacteriawasaccompaniedbyprolonged group B were killed at serial time intervals with six rats
survivalcomparedwithseveresepsiscausedbysuscepti- per interval.Atevery timeofsacrifice, amidline abdom-
ble bacteria [11]. As a consequence, it may be hypothe- inal incisionwasperformed. Intestineswere displacedto
sizedthatthepathogenesisofinfectionsbyMDRbacteria the left, and the inferior vena cava was recognized and
mayhavespecificpeculiarities. puncturedwitha19-gaugeneedle.Fivemillilitersofblood
It is thus evident that in-depth understanding is was collected into pyrogen-free syringes and applied for
required on the process of lipid peroxidation for infec- themeasurementoftumornecrosisfactor-alpha(TNF-a)
tions caused by MDR bacteria. This is based on (a) the andofmalondialdehyde(MDA).Animalswerethenkilled
compartmentalizationoftheprocessoflipidperoxidation withanintramuscularinjectionofpentothal.
for the generation of organ failures coming from one These experiments were repeated in six animals per
cohortofpatientswithseveresepsis[7];and(b)different group at the defined peak time point of serum MDA.
clinical characteristics of severe sepsis elicited by MDR Understerileconditions,segmentsfromtheliver,spleen,
bacteria from those by susceptible bacteria [11]. To this lower lobe of the right lung, the heart, the right kidney
end,adualapproachwasused:first,westudiedlipidper- and of the abdominal aorta were taken and placed into
oxidation in tissues in a lethal sepsis model with MDR separate sterile plastic containers. In these experiments,
Toufekoulaetal.CriticalCare2013,17:R6 Page3of11
http://ccforum.com/content/17/1/R6
Figure1Studyflowchart.MDA,malondialdehyde;MDR,multidrugresistant;VAP,ventilator-associatedpneumonia.Thechestradiographisof
apatientenrolledinthestudycohort,withpermission.
5 ml of blood was sampled after venipuncture of the removed and incubated with 2 ml of thiobarbituric acid,
lower vena cava under aseptic conditions and collected 0.2% (Merck) for 60 minutes at 90°C. After centrifuga-
into EDTA-containing tubes (Becton Dickinson, Cock- tion, a volume of 10 μl of the supernatant was injected
eysville,MD,USA)forthemeasurementwithflow cyto- into a high-performance liquid chromatography system
metry of an oxidative burst on neutrophils. Tissue (HPLC; Agilent 1100 Series, Waldbronn, Germany) with
segmentswere weighedandhomogenized by using1 ml the following characteristics of elution: Zorbax Eclipse
of sterile phosphate-buffered saline (PBS; pH 7.2). For XDB-C18 (4.6 × 150 mm, 5 μm) column under 37°C;
the measurement of tissue bacterial outgrowth, aliquots mobile phase consisting of 50 mM K PO (pH 6.8) buf-
3 4
of0.1mlofthehomogenateswerediluted1:10intoster- fer, and methanol 99% at a 60:40 ratio with a flow rate
ile 0.9%NaClforfourconsecutivetimes.Aliquotsof0.1 of 1 ml/min, fluorometric detection with signals of exci-
mlofeachdilutionwereplatedontoMcConkeyagarand tation at 515 nm and emission at 535 nm. The retention
incubatedat35°Cfor3days.Thenumberofviablecolo- time of MDA was 3.5 minutes, and it was estimated in
nies was counted by multiplying with the appropriate millimoles per milliliter with a standard curve created
dilutionfactor.Thelowerlimitofdetectionwas10cfu/g. with 1,1,3,3-tetramethoxy-propane (Merck). All determi-
The number of viable cells in tissues was expressed by nations were performed in duplicate. The lowest limit of
log values. detection was 0.1 mmol/ml, and the interday variation
10
Lipid peroxidation was estimated in serum and in tis- of the assay was 1.01%. Tissue concentrations of MDA
sue homogenates by the concentration of MDA, as were expressed as millimoles per gram.
already described [12]. In brief, a 0.1-ml aliquot of each Concentrations of serum TNF-a were measured in
sample was mixed with 0.9 ml of trichloroacetic acid duplicate with an enzyme immunoassay (Diaclone,
20% (Merck, Darmstadt, Germany) and centrifuged at Marseille, France). The lower limit of detection was
12,000 g and 4°C for 10 minutes. The supernatant was 12 pg/ml.
Toufekoulaetal.CriticalCare2013,17:R6 Page4of11
http://ccforum.com/content/17/1/R6
For themeasurement oftheoxidativeburstonneutro- Statistical analysis
phils, 0.1 ml of EDTA-blood was aliquoted into separate For animal experiments, survival was estimated with
sterile tubes. A volume of 20 μl of phorbol-myristate Kaplan-Meieranalysis.ConcentrationsofMDAandTNF-
(PMA, Sigma, St. Louis, MO, USA; final concentration, awereexpressedbytheirmeans±SEM.Comparisonsof
10 μΜ) was added in the second tube, and the first tube concentrationsofMDAandofTNF-ainserumbetween
wasleftuntreated.Bothtubeswereincubatedfor10min- serialtimesofsamplingweredonewithone-way analysis
utes at 37°C in a water bath. Then 10 μl of DHR (dihy- of variance (ANOVA) with post hoc Bonferroni correc-
drorhodamine; Sigma, final concentration 100 μΜ) was tions.ComparisonsoftissueMDAbetweengroupsAand
added. After10 minutes of incubationat 37°Cinawater BwereperformedwiththeMann-WhitneyUtest.Corre-
bath, 1 ml of VersaLyse was added in both tubes for the lations between tissue MDA and bacterial growth were
lysis of red blood cells (Immunotech, Marseille, France). accordingtoSpearman’srankoforder.
After 15 minutes of incubation in the dark, tubes were Fortheclinicalstudy,theMDAofthefirstdayinserum
analyzed with the FC500 counter (Beckman Coulter, was expressed as median and 95% confidence intervals.
Miami, FL, USA). Oxidative burst was expressed as the Comparisons between patients with organ failures and
mean fluorescence intensity (MFI) of DHR on granulo- those without any organ failures were done with the
cytesaftergatingbythecharacteristicFS/SSscatteringby Mann-Whitney U test, analyzing also for outliers. To
usingPMA-untreatedcellsasnegativecontrols. demonstratetheimpactoflipidperoxidationinthepatho-
genesis of liver dysfunction and of renal dysfunction, all
Clinical study patients with acute hepatic dysfunction and all patients
Serum samples were collected daily on the first 7 days with acute renal dysfunction were grouped together as
from93patientsenrolledintheplaceboarmofadouble- havinghepaticdysfunctionandrenaldysfunction,respec-
blind,randomized,placebo-controlledclinicaltrialforthe tively, whether or not they had other organ failures.
managementofVAPandsepsis[13].Thestudywasper- MDAs of consecutive days were expressed as mean ±
formed during the period from June 2004 to November SEM. Comparisons between survivors and nonsurvivors
2005,andpatientswerehospitalizedintheDepartmentof weredonewiththeMann-WhitneyUtest.Forallcompar-
Critical Care Medicine of the “Evangelismos” General isons,anyvalueofP<0.05 afteradjustmentformultiple
Hospital,inthe2ndDepartmentofCriticalCareMedicine comparisonswithBonferroniwasconsideredsignificant.
ofthe“ATTIKON”University Hospital ofAthens andin
the 4th Department of Internal Medicine of the “ATTI- Results
KON”UniversityHospitalofAthens.Thestudywascon- Animal infection model
ducted after approval of the Ethics Committees of Mortality ofthe studied sepsismodel by MDRP. aerugi-
EvangelismosGeneralHospitalandofATTIKONUniver- nosa was 50% at 12 hours after bacterial challenge and
sity Hospital andafter writteninformedconsentbyfirst- 100% at 36 hours after bacterial challenge. None of the
degree relatives. Patients were divided into those with sham-operated-ratsdied.Sacrificeexperimentsconducted
hepatic dysfunction, acute respiratory distress syndrome atserialtimeintervalsafterbacterialchallengefoundthe
(ARDS),acuterenal dysfunction, andcardiovascularfail- peakofserumMDAat5hours(Figure2).Thiscoexisted
ure(CV)bystandarddefinitions[14].Threemicrolitersof withthepeakoftheproinflammatoryresponseasindexed
bloodwassampledfor7consecutivedaysaftervenipunc- by serum TNF-a. Based on this finding, it was decided
tureofoneforearmveinunderasepticconditionsandcol- that sacrifice experiments with measurement of tissue
lected into pyrogen-free tubes (Vacutainer, Becton lipid peroxidation should be repeated in rats killed at
Dickinson).Thetubeswereimmediatelycentrifuged,and 5hoursafterchallengewiththetestisolate.
serumwas kept refrigerated at -70°Cuntil assayed. Con- These experiments showed that MDA was particularly
centrationsofMDAweremeasuredinserum,asdescribed elevatedinliver,inspleen,andintheaorticwall(Figure3);
earlier. however,MDAwassignificantlyreducedintherightkid-
ney.PairedcomparisonsbytheWilcoxonsigned-ranktest
Study end points showedthatMDAintheliverwasgreaterthanthatinthe
Theprimarystudyendpointwastheexistenceofcompart- lung (P = 0.016), in the spleen (P = 0.041), in the heart
mentalizationoflipidperoxidationinrelationtothespeci- (P =0.004), but not in the aortic wall (P = 0.116). At the
fic failing organs. This was assessed in an experimental timeofdeath,tissuesofsham-operated-onratsweresterile.
modelinrodents,anditwastestedforconfirmationbythe However,mean±SEMcountsoftheinjectedisolateinthe
measurementofcirculatingMDAinpatientswithsepsis. liver were log 3.95 ± 0.63 cfu/g; in the spleen, log
10 10
The secondary study end point was the correlation 4.29±0.71cfu/g;inthelung,log 3.07±0.76cfu/g;inthe
10
between circulating MDA and final outcome in relation heart,log 2.55±0.58cfu/g;andintheaorticwall,log
10 10
to the underlying organ failure. 3.88 ± 0.67 cfu/g. Paired comparisons did not show any
Toufekoulaetal.CriticalCare2013,17:R6 Page5of11
http://ccforum.com/content/17/1/R6
is also the situation for human sepsis, we measured
circulating MDA in 93 patients with sepsis, making the
hypothesis that serum MDA should differ between
patients as a reflection of the failing organ. These
patients were enrolled in the placebo arm of a rando-
mized trial previously published [13,15]. All had VAP,
and the pathogens isolated in the quantitative tracheo-
bronchial secretions were MDR species of Acinetobacter
baumannii, of Pseudomonas aeruginosa, and of Kleb-
siella pneumoniae. As a consequence, this study cohort
could be used to extrapolate animal findings because all
patients had the same source of sepsis (that is, VAP),
and they were infected with MDR bacteria. The demo-
graphic characteristics of these patients were described
elsewhere [13].
Analysis of circulating MDA was done in relation to
the type of failing organ. From these 93 patients, 18 did
nothaveanyorganfailure;themean±SEMpO /FiO of
2 2
the first day of sampling of these patients was 338.9 ±
34.8 mm Hg. This pO /FiO ratio showed that they did
2 2
not have acute lung injury, and they could be used as
appropriate comparators for this analysis. Another 10
patients had hepatic dysfunction, 23had only ARDS; 25
had ARDS and CV failure; 14 had CV failure without
ARDS; and 10 had acute renal dysfunction. Of the 10
patients with acute hepatic dysfunction, four also had
other organ failures. Of the 10 patients with acute renal
dysfunction,threealsohadotherorganfailures.Circulat-
ingconcentrationsofMDAareshowninFigure5.Itwas
obvious that circulating MDA was increased in the case
of hepatic dysfunction and in the case of ARDS. In
Figure2Serialconcentrationsoftumor-necrosisfactor-a(TNF- patientswithbothARDSandCVfailureorwithCVfail-
a)andmalondialdehyde(MDA)afterbacterialchallenge.
ure without ARDS, MDA did not increase. Circulating
Experimentalsepsiswasinducedinratsaftertheintraperitoneal
MDAwaslower inpatientswithacuterenal dysfunction
injectionofonemultidrug-resistantisolateofPseudomonas
aeruginosa.Attheindicatedtimeintervalsafterchallenge(n=6per thaninpatientswithoutorgan failuresandcorroborated
time),ratswerekilled,andconcentrationsofMDA(a)andofTNF-a thefindingsoftheexperimentalinfectionmodelinrats.
(b)weremeasuredinserum.*P<0.05comparedwiththe5-hour Of the 93 enrolled patients, the implicated pathogen
intervalafteradjustmentformultiplecomparisons.
was isolated from quantitative tracheobronchial secre-
tions at a quantity > 105 cfu/ml from 60 patients. More
precisely, Pseudomonas aeruginosa was isolated from 10
difference betweenthebacterialgrowthofthe livercom- patients, Acinetobacter baumannii from 38 patients,
pared with the other tissues. No significant correlation Klebsiella pneumoniae from six patients, and other
couldbefoundbetweenMDAandbacterialgrowthintis- gram-negative species from six patients. Microbiology
sues(correlation forliver,r ,-0.273;P=0.417;forlung, documentation failed in 30 patients. Serum MDA in
est
r ,+0.040;P=0.841;forspleen,r,-0.445;P=0.170;for relation with the microbiology of VAP is shown in
est s
rightkidney,r,+0.080,P=0.815;forheart,r:-0.516,P= Figure 6. No differences were found in relation to the
s s
0.295;andforaorta,r:+0.123,P=0.715;datanotshown). underlying pathogen (P = 0.925).
s
At the time of death, the neutrophil burst was defec- Follow-upmeasurementsinsurvivorsandnonsurvivors
tive in circulating neutrophils of infected rats (Figure 4). fromorgan failuresareshown inFigure7.Time-kinetics
of MDA differed in relation to the type of failing organ
Clinical study (s).Moreprecisely,serumMDAwasgreaterinsurvivors
Results of the animal experiments prompted us to fromhepaticdysfunctiononday4andinsurvivorsfrom
hypothesize that a compartmentalization of the oxidant ARDS and CV failure on day 5. Serum MDA was also
status takes place in sepsis. To investigate whether this lower on day 4 among survivors from acute renal
Toufekoulaetal.CriticalCare2013,17:R6 Page6of11
http://ccforum.com/content/17/1/R6
Figure3TissueconcentrationsofMDAafterbacterialchallenge.Experimentalsepsiswasinducedinratsaftertheintraperitonealinjection
of2×107cfu/animalofonemultidrug-resistantisolateofPseudomonasaeruginosa.Fivehoursafterchallenge(n=6pergroup),ratswere
killed,andconcentrationsofmalondialdehyde(MDA)weremeasuredinthetissuehomogenatesofliver(a),lung(b),spleen(c),rightkidney(d),
heart(e),andaorta(f).Pvaluescomparedwithsham-operatedratsareshown.
dysfunction.Similarstatisticalcomparisonscouldnotbe kidneys. It should be overemphasized that, at the time
doneforpatientswithsingleARDS,becausethenumber of animal death, when lipid peroxidation differed con-
ofnonsurvivorswaslow(n=2). siderably within tissues, the oxidative burst of circulating
No statistical correlation could be found between neutrophils was defective. This finding is in accordance
serum MDA and survival time for any of these sub- with reports that the function of neutrophils becomes
groups of patients (data not shown). defective in severe sepsis in exactly the same way that
immunoparalysis succumbs in the constellation of infec-
Discussion tion-induced organ failures [17]. However, this finding
Excess formation of lipid peroxides in serum accompa- may also reflect, at least in part, that in the studied set-
nied by consumption of the total antioxidant capacity of ting of widespread systemic infection by MDR P. aerugi-
blood has been demonstrated to take place over the nosa, oxidative stress fails in circulating neutrophils, but
time-course of experimental bacteremia by MDR P. aer- it is highly generated in tissues, probably as a result of
uginosa in rabbits [16]. In the present study, in-depth infiltration by neutrophils.
investigation in an animal model of lethal sepsis by the These findings corroborate many of the findings of
same MDR species revealed that lipid peroxidation was other investigators, but they also present discrepancies
highly compartmentalized. Excess formation of MDA withotherstudiesinrodents.Inthesestudies, investiga-
took place in the liver, in the spleen, and in the aortic tors used either purified bacterial lipopolysaccharide
wall, whereas that was not the case for the lung and for (LPS) or CLP to stimulate host defenses. MDA in liver
the heart. On the contrary, MDA was decreased in the homogenates had been increased in rats killed 6 hours
Toufekoulaetal.CriticalCare2013,17:R6 Page7of11
http://ccforum.com/content/17/1/R6
The human situation is far more difficult; comorbid
conditions play a pivotal role, and only indirect simula-
tions can be done. Trying to validate the generated ani-
mal data in human sepsis, it was decided that one
hospital-acquired infection by MDR gram-negative bac-
teria should be studied. To this end, a cohort of patients
with VAP by MDR gram-negative species was studied
[13]. This cohort proved to be an adequate equivalent of
clinical sepsis by MDR bacteria, because circulating
MDA did not differ in relation to the implicated bacter-
ial species.
Circulating MDA was correlated with the types of fail-
ing organs. Results indicated that MDA was particularly
elevated in specific organ failures compared with
patients with none organ failure corroborating the ani-
mal data for compartmentalization of lipid peroxidation.
Figure4Oxidativeburstofcirculatingneutrophils.Experimental
In a recent study [7], concentrations of the lipid meta-
sepsiswasinducedinratsaftertheintraperitonealinjectionofone
bolites F2-isoprostane and isoflurane were measured in
multidrug-resistantisolateofPseudomonasaeruginosa.Fivehours
afterchallenge(n=6pertime),ratswerekilled,andoxidativeburst 50 patients with severe sepsis. These metabolites were
onneutrophilswasmeasuredastheexpressionof considered an indirect index of lipid peroxidation. Mea-
dihydrorhodamine(DHR)oncellmembranesafterflow-cytometric sured concentrations were greater in patients who devel-
analysis.Pvaluecomparedwithsham-operatedratsisshown.
oped hepatic failure compared with those who did not.
Contrary to the presented clinical findings, F2-isopros-
tane and isoflurane were greater in the event of renal
after the intraperitoneal injection of LPS [18]. When a failure. In the same study of 50 patients with severe sep-
modelofCLPwasstudiedinrats,downregulationofthe sis [7], levels of lipid metabolites remained unchanged
expressionofgenescoupledwithmitochondrialfunction in the event of ARDS and of CV failure. Many studies
was found [19]. This is consistent with mitochondrial have provided evidence for excess lipid formation in the
uncouplingfromexcessROSformationandcorroborates event of acute lung injury and of ARDS [23]. This is
the presented findings. However, the described lack of probably related to the high infiltration of the lung by
excess lipid peroxidation in the heart muscle after chal- neutrophils that overproduce ROS. Our study indicated
lenge by MDR P. aeruginosa herein is in disagreement that ARDS ensuing in the absence of CV failure is
with two recent publications. In the first publication in accompanied by excess lipid peroxidation. However,
mice [20], excessROS formationwasfound inthe myo- when ARDS is accompanied by CV failure, excess circu-
cardium48hoursafterCLP.Inthepresentstudy,animal lating MDA ceases to exist. The study of the levels of
killing was done at 5 hours (that is, earlier than the F2-isoprostane and isoflurane in severe sepsis did not
mouse CLP model), and this may explain the discre- study separately patients with ARDS without CV failure
pancy. In the second publication, excess formation of from patients with both ARDS and CV failure.
ROS was found in the heart tissue as early as 3 hours Many of the conducted clinical studies on the forma-
after bacterial challenge.However,thiswasacompletely tion of ROS in human sepsis have shown that excess
different model of sepsis induced after aspiration chal- lipid peroxidation is an index of unfavorable outcome
lengewithStreptococcuspneumoniae[21]. [5,6,23]. Findings of the present study suggest that
Infections by MDR bacteria emerge as a serious pro- kinetics of MDA in serum differ between survivors and
blem in the hospital environment. They colonize nonsurvivors, depending on the type of organ failure.
patients after long-term hospital stay, and they are the More precisely, MDA is increased over the disease
cause of severe sepsis after ICU admission. Available course of survivors of hepatic dysfunction and of ARDS.
choices for antimicrobial therapy become more and On the contrary, it is decreased over the time course of
more limited, creating major hurdles. Limitations of survivors of acute renal dysfunction. The results of the
management shed light on interventions on the host current study do not provide mechanistic data to
response. Antioxidant supplementation against excess explain the findings, and only assumptions can be made.
formation of ROS and RNS belongs to these interven- Regarding the increase of serum MDA within nonsurvi-
tions [22]. vors of acute renal dysfunction, it may be hypothesized
Animal sepsis models offer the advantage of well- that although lipid peroxidation in not the principal
defined infection settings and of clear-cut approaches. mechanism for acute renal dysfunction, it participates in
Toufekoulaetal.CriticalCare2013,17:R6 Page8of11
http://ccforum.com/content/17/1/R6
Figure5SerumconcentrationsofMDAinpatientswithsepsis.Concentrationsofmalondialdehyde(MDA)weremeasuredonthefirstday
ofsepsisduetoventilator-associatedpneumoniawithmultidrug-resistantgram-negativebacteriain93patients.Resultsarepresentedaccording
tothetypesoffailingorgans:(a)patientswithhepaticdysfunctionversuspatientswithoutanyorganfailure;(b)patientswithARDSversus
patientswithoutanyorganfailure;(c)patientswithARDSandCVfailureversuspatientswithoutanyorganfailure;(d)patientswithCVfailure
withoutARDSversuspatientswithoutanyorganfailure;and(e)patientswithacuterenaldysfunctionversuspatientswithoutanyorganfailure).
Circlesdenoteoutliers.Pvaluesreflectcomparisonsbetweenpatientswithoutanyorganfailureandwiththeindicatedorganfailure.ARDS,
acuterespiratorydistresssyndrome;CV,cardiovascular.
the progression toward unfavorable outcome. Further- andphagocytosis,whereasthesesamefunctionsareupre-
more, different patterns of the kinetics of serum MDA gulatedinperipheralorganslikethespleenandthelung.
are shown in survivors and nonsurvivors of CV failure The complementary findings between the described
with and without ARDS. Taking into consideration the animal model and human sepsis are highly indicative of
respective patterns in patients only with ARDS, it may an existing compartmentalization of lipid peroxidation
be postulated that overall, serum MDA is elevated in in systemic infections with MDR gram-negative bacteria.
the event of ARDS; this is decreased on addition of CV It should be underscored that such a study design is not
failure and further decreased in those only with CV devoid of limitations. The major limitation is the time
failure. course until the start of one organ failure, which is
Animal sepsis with MDR P. aeruginosa was accompa- determined by the investigators of an animal study, in
niedbya defective oxidative burstof circulatingneutro- contrast to the clinical situation in which this is not
phils.However,tissueMDAofthespleenwasincreased. possible to be defined with precision. However, the
These findings are in accordance with the decrease in described compartmentalization may explain, at least in
serum activity of myeloperoxidase described in patients part, the contradictory findings of clinical studies with
with septic shock[24]. It may thus be hypothesizedthat antioxidant supplementation in severe sepsis [23].
in theeventof severe sepsisand organ failures,circulat- Results of the present study can be used to select the
ingneutrophilsare defectiveinefficientoxidative bursts populations of patients at severe sepsis that may benefit
Toufekoulaetal.CriticalCare2013,17:R6 Page9of11
http://ccforum.com/content/17/1/R6
from antioxidant agents, with emphasis on patients with
hospital-acquired sepsis with MDR gram-negative spe-
cies. Although it cannot be precluded that lipid peroxi-
dation may be stimulated in a similar mechanism by
gram-negative bacteria and by gram-positive cocci, the
present results are limited on gram-negative sepsis, par-
ticularly because gram-negative species emerged as the
main pathogens of the studied cohort.
Conclusions
These findings coming from an experimental infection
model that are confirmed from a patient cohort clearly
indicate that, during systemic infections, lipid peroxida-
tion is characterized by high compartmentalization. The
phenomenon is considerably upregulated in the event of
ARDS and hepatic dysfunction and downregulated in
Figure6SerumconcentrationsofMDAinpatientswithsepsis.
Concentrationsofmalondialdehyde(MDA)weremeasuredonthe the event of acute renal dysfunction.
firstdayofsepsisduetoventilator-associatedpneumoniawith
multidrug-resistantgram-negativebacteriain93patients.Resultsare Key messages
presentedaccordingtothetypeofimplicatedpathogen.Circles (cid:129) High compartmentalization of lipid peroxidation
denoteoutliers,andasterisksdenoteextremes.
exists in experimental P. aeruginosa infection
Figure7Follow-upmeasurementsofMDAofseruminpatientswithsepsis.Concentrationsofmalondialdehyde(MDA)weremeasuredfor
7consecutivedaysinpatientswithventilator-associatedpneumoniawithmultidrug-resistantgram-negativebacteriaandsepsis.Resultsare
presentedaccordingtothetypeoffailingorgan:(a)patientswithhepaticdysfunction;(b)onlyARDSpatients;(c)patientswithARDSandCV
failure;(d)patientswithCVfailurewithoutARDS;and(e)patientswithacuterenaldysfunction,andseparatelyforsurvivors(S)andnonsurvivors
(NS).*P<0.05betweenSandNS.
Toufekoulaetal.CriticalCare2013,17:R6 Page10of11
http://ccforum.com/content/17/1/R6
(cid:129) Increased levels are found in liver, spleen, and RoussouZ,ClouvaP,MaguinaN,KanellakopoulouK,ArmaganidisA,
aorta, and decreased levels are found in kidney. GiamarellouH:Tigecyclineinthetreatmentofinfectionsfrommulti-
drugresistantgram-negativepathogens.JInfect2009,58:273-284.
(cid:129) Results are confirmed in a cohort of patients with
9. KangCI,KimSH,ParkWB,LeeKD,KimHB,KimEC,OhMD,ChoeKW:
gram-negative sepsis Bloodstreaminfectionscausedbyantibiotic-resistantgram-negative
(cid:129) Levels of MDA are increased in hepatic dysfunc- bacilli:riskfactorsformortalityandimpactofinappropriateinitial
antimicrobialtherapyandoutcome.AntimicrobAgentsChemother2005,
tion and ARDS
49:760-766.
(cid:129) Levels of MDA are decreased in renal dysfunction 10. TeoJ,CaiY,TangS,LeeW,TanTY,TanTT,KwaAL:Riskfactors,molecular
and remain unchanged in cardiovascular failure epidemiologyandoutcomesofertapenem-resistant,carbapenem-
susceptibleEnterobacteriaceae:acase-case-controlstudy.PLoSOne2012,
7:e34254.
11. Giamarellos-BourboulisEJ,PapadimitriouE,GalanakisN,AntonopoulouA,
Abbreviations TsaganosT,KanellakopoulouK,GiamarellouH:Multidrug-resistanceto
ARDS:acuterespiratorydistresssyndrome;CLP:cecalligationandpuncture; antimicrobialsasapredominantfactorinfluencingpatientsurvival.IntJ
CV:cardiovascular;DHR:dihydrorhodamine;LPS:lipopolysaccharide;HPLC: AntimicrobAgents2006,27:476-481.
high-performanceliquidchromatography;MDA:malondialdehyde;MDR: 12. AgarwalR,ChaseSD:Rapid,fluorometric-liquidchromatographic
multidrugresistant;PMA:phorbol-myristateacetate;ROS:reactiveoxygen determinationofmalondialdehydeinbiologicalsamples.JChromatogr
species;TNF-α:tumornecrosisfactor-alpha;VAP:ventilator-associated 2002,775:121-126.
pneumonia. 13. Giamarellos-BourboulisEJ,PechèreJC,RoutsiC,PlachourasD,KolliasS,
RaftogiannisM,ZervakisD,BaziakaF,KoronaiosA,AntonopoulouA,
Authors’contributions MarkakiV,KoutoukasP,PapadomichelakisE,TsaganosT,ArmaganidisA,
CTperformedmeasurementsofMDA,draftedthemanuscript,andagreedto KoussoulasV,KotanidouA,RoussosC,GiamarellouH:Effectof
thefinalsubmittedversion.VP,TT,andDPCperformedanimalexperiments, clarithromycininpatientswithsepsisandventilator-associated
draftedthemanuscript,andagreedtothefinalsubmittedversion.CR,SEO, pneumonia.ClinInfectDis2008,46:1157-1164.
AK,MR,andFBenrolledpatients,collectedserumsamplesandclinicaldata, 14. LevyM,FinkMP,MarshallJC,AbrahamE,AngusD,CookD,CohenJ,
draftedthemanuscript,andagreedtothefinalsubmittedversion.EJGB OpalSM,VincentJL,RamsayG:2001SCCM/ESICM/ACCP/ATS/SIS
designedthestudy,wrotethemanuscript,andagreedtothefinal internationalsepsisdefinitionsconference.CritCareMed2003,
submittedversion. 31:1250-1256.
15. SpyridakiA,RaftogiannisM,AntonopoulouA,TsaganosT,RoutsiC,
Competinginterests BaziakaF,KaragianniM,MouktaroudiM,KoutoukasP,PelekanouA,
Theauthorsdeclarethattheyhavenocompetinginterests. KotanidouA,OrfanosSE,vanderMeerJWM,NeteaMG,Giamarellos-
BourboulisEJ:Effectofclarithromycinininflammatorymarkersof
Acknowledgements patientswithventilator-associatedpneumoniaandsepsiscausedby
ThestudywassupportedinpartbybytheFederalMinistryodEducation Gram-negativebacteria:resultsfromarandomizedclinicalstudy.
andResearch(BMBF),Germany(FKZ:01EO1002) AntimicrobAgentsChemother2012,56:3819-3825.
16. KoussoulasV,Giamarellos-BourboulisEJ,AdamisT,MouktaroudiM,
Authordetails SabracosL,PerreaD,GiamarellouH,Dionyssiou-AsteriouA:The
14thDepartmentofInternalMedicine,UniversityofAthens,MedicalSchool, significanceofoxidant/antioxidantbalanceforthepathogenesisof
12462Athens,Greece.21stDepartmentofCriticalCareMedicine,University experimentalsepsisbymultidrug-resistantPseudomonasaeruginosa.
ofAthens,MedicalSchool,10676Athens,Greece.32ndDepartmentofCritical ProstaglandinsLeukotEssentFattyAcids2005,72:41-47.
CareMedicine,UniversityofAthens,MedicalSchool,12462Athens,Greece. 17. Alves-FilhoJC,SpillerF,CunhaFQ:Neutrophilsparalysisinsepsis.Shock
4IntegratedResearchandTreatmentCenter,CenterforSepsisControland 2010,34(Suppl1):15-21.
Care(CSCC),JenaUniversityHospital,07747Jena,Germany. 18. MohamandinAM,ElberryAA,ElkablawyMA,GawadHAS,Al-AbassiFA:
Monelukast,aleukotrienereceptorantagonistabrogates
Received:12October2012 Revised:20December2012 lipopolysaccharide-inducedtoxicityandoxidativestressinratliver.
Accepted:9January2013 Published:16January2013 Pathophysiology2011,18:235-242.
19. YangQ,MattickJSA,OrmanMA,NguyenTT,IerapetritouMG,BerthiaumeF,
References AndroulakisIP:Dynamicsofhepaticgeneexpressionprofileinaratcecal
1. BelikJ,González-LuisG,Perez-VizcainoF,VillamorE:Isoprostanesinfetal ligationandpuncturemodel.JSurgRes2012,176:583-600.
andneonatalhealthanddisease.FreeRadicalBiolMed2010,48:177-180. 20. SmedingL,Leong-PoiH,HuP,ShahY,HaitsmaJJ,HorvathE,FurmilS,
2. UchidaK:4-Hydroxy-2nonenal:aproductandmediatorofoxidative MasoomH,KuiperJW,SlutskyAS,ParkerFB,dosSantosCC:Salutary
stress.ProgLipidRes2003,42:318-343. effectsofresveratrolonsepsis-inducedmyocardialdepression.CritCare
3. AzenaborAA,MahonyJB:Generationofreactiveoxygenspeciesand Med2012,40:1896-1907.
formationofmembranelipidperoxidesincellsinfectedwithChlamydia 21. ZhangQS,SadekH,MaassDL,MartinezB,MaL,KilgoreJA,WilliamsNS,
trachomatis.IntJInfectDis2000,4:46-50. FrantzDE,WiggintonJG,NwariakuFE,WolfSE,MineiJP:Specificinhibition
4. GalleyHF:Oxidativestressandmitochondrialdysfunctioninsepsis.BrJ ofmitochondrialoxidativestresssuppressesinflammationandimproves
Anesth2011,107:57-64. cardiacfunctioninratpneumonia-relatedsepsismodel.AmJPhysiol
5. GelaimDP,deBittencourtPasqualMA,ComimCM,GrunwaldMS,RitterC, HeartCircPhysiol2012,302:H1847-H1859.
DamianiTomasiC,CascaesAlvesS,QuevedoJ,Dal-PizzolF,Donsace 22. AngstwurmMW,EngelmannL,ZimmermannT,LehmannC,SpesCH,
MoreiraJC:Serumheat-shockprotein70levels,oxidantstatus,and AbelP,StraussR,Meier-HellmannA,InselR,RadkeJ,SchüttlerJ,GärtnerR:
mortalityinsepsis.Shock2011,35:466-470. Seleniuminintensivecare(SIC):resultsofaprospectiverandomized,
6. AndersenM,RegueiraT,BruhnA,PerezD,StrobelP,DougnacA, placebo-controlled,multiple-centerstudyinpatientswithsevere
MarshallG,LeightonF:Lipoperoxidationandproteinoxidativedamage systemicinflammatoryresponsesyndrome,sepsis,andsepticshock.Crit
exhibitdifferentkineticsduringsepticshock.MediatorsInflamm2008, CareMed2007,35:118-126.
168652. 23. CrimiE,SicaV,Williams-IgnaroS,ZhangH,SlutskyAS,IgnarroLJ,NapoliC:
7. WareLB,FesselJP,MayAK,RobertsLJII:Plasmabiomarkersofoxidant Theroleofoxidativestressinadultcriticalcare.FreeRadicBiolMed2006,
stressanddevelopmentoforganfailuresinseveresepsis.Shock2011, 40:398-406.
36:12-17. 24. KothariN,KeshariRS,BograJ,KohliM,AbbasH,MalikA,DikshitM,
8. PoulakouG,KontopidouF,ParamythiotouE,KompotiM,KatsiariM, BarthwalMK:Increasedmyeloperoxidaseenzymeactivityinplasmaisan
MainasE,NicolaouC,YphantisD,AntoniadouA,Trikka-GraphakosE,