Review
Reactive Oxygen Comes of Age:
Mechanism-Based Therapy of Diabetic End- Organ Damage
Mahmoud H. Elbatreek,
1,2,* Mayra P. Pachado,
1Antonio Cuadrado,
3Karin Jandeleit-Dahm,
4and Harald H.H.W. Schmidt ,
1,*
Reactive oxygen species (ROS) have been mainly viewed as unwanted by- products of cellular metabolism, oxidative stress, a sign ofa cellular redox imbalance, and potential disease mechanisms, such as in diabetes mellitus (DM).Antioxidant therapies, however, have failedto provide clinical benefit.
Thisparadox can be explained byrecent discoveries thatROS have mainly essentialsignalingandmetabolicfunctionsandevolutionallyconservedphys- iological enzymatic sources. Disease can occur when ROS accumulate in nonphysiological concentrations, locations, or forms. By focusing on dis- ease-relevantsourcesandtargetsofROS,andleavingROSphysiologyintact, precise therapeutic interventions are now possible and are entering clinical trials.TheiroutcomesarelikelytoprofoundlychangeourconceptsofROSin DMandinmedicineingeneral.
ANewApproachtoDiabetesMellitusand ReactiveOxygenSpecies
Diabetesmellitus (DM) andits related end-organ damage, such as diabetic nephropathy, neuropathy, retinopathy, and cardiomyopathy, are major causes of death and long-term disability.Their underlyingmechanismsare incompletelyunderstood,whichiswhynoneof thecurrentantidiabetictherapiestargettheunderlyingcausesorarecurative,butfocusinstead onnormalizingsurrogateparametersorriskfactorssuchasbloodglucoseorhypertension[1].
Hence,ourlackofmechanisticunderstandingoflifestylechange-resistantdiabeticend-organ damage,togetherwiththeincreasingprevalenceofDM,representasignificantmajorunmet medicalneed.
Onemechanismthathasbeensuggestedfordecadestocausepancreaticbcelldysfunction anddiabeticend-organdamageis‘oxidativestress’(seeGlossary),originallydefinedasan overproductionofreactiveoxygenspecies(ROS).Antioxidantswereconsideredtheobvi- oustherapeuticcountermeasurebut,clinically,haveconsistentlydisappointed[2].Evenworse, meta-analysesofclinicaltrialsshowthatantioxidantsmaynotonlybeineffective,butharmful, andevenincreasemortality[2].
Recently, however, important conceptual breakthroughs in our understanding of ROS in general and DM in particular explain the failure of antioxidants and point towards entirely differentmechanism-basedandpossiblycurative therapeuticapproaches.Our newunder- standingofROSrequiresthatmanylong-heldmisconceptions,suchasthe‘redoxbalance hypothesis’andtheviewthatROSareprimarilystressors,diseasetriggers,andmetabolic wasteproducts,mustbeovercome.Instead,the manyphysiologicalrolesofROSandthe existence of at least seven evolutionarily conserved ROS-producing enzymes (NOX1–5,
Highlights
Lackofcomprehensiveunderstanding ofDManditscomplications,together withitsincreasingprevalence,repre- sent a major unmet medical need.
Thereisquitealotknownaboutits mechanisms,butmoreisneeded.
Reactiveoxygenspecies(ROS) may hold the answer to this need, but long-standing misconceptions about oxidativestressandtheinefficacyof antioxidants have prevented clinical breakthroughs.
AnewunderstandingofROSasboth essential signaling molecules and pathomechanismhaveledtotheiden- tification, at different stages ofdia- betes, of different therapeutically relevantsourcesandtargetsofROS.
Networkpharmacologyandprecision diagnostics, currently in advanced stagesofclinicaldevelopment,incon- junction with essential lifestyle changes,willrevolutionizethetherapy andpreventionofdiabetesanditsend- organdamage.
1DepartmentofPharmacologyand PersonalisedMedicine,Facultyof Health,MedicineandLifeSciences, MaastrichtUniversity,Maastricht,The Netherlands
2DepartmentofPharmacologyand Toxicology,FacultyofPharmacy, ZagazigUniversity,Zagazig,Egypt
3DepartmentofBiochemistry,Faculty ofMedicine,AutonomousUniversity ofMadrid,InstitutodeInvestigaciones BiomédicasUAM-CSIC,Cibersobre EnfermedadesNeurodegenerativas (CIBERNED),InstitutodeInvestigación
312 TrendsinEndocrinology&Metabolism,May2019,Vol.30,No.5 https://doi.org/10.1016/j.tem.2019.02.006
©2019TheAuthors.PublishedbyElsevierLtd.ThisisanopenaccessarticleundertheCCBY-NC-NDlicense(http://creativecommons.org/licenses/by-nc- nd/4.0/).
SanitariaLaPaz(IdiPaz),Madrid, Spain
4DepartmentofDiabetes,Central ClinicalSchool,MonashUniversity, Melbourne,Australia
*Correspondence:
m.elbatrik@maastrichtuniversity.nl(M.
H.Elbatreek)and
h.schmidt@maastrichtuniversity.nl (HaraldH.H.W.Schmidt).
DUOX1–2),inadditiontotheiralternativelysplicedvariants,shouldberecognized.Inparticular, thediscoveryofNADPHoxidaseisoforms,whoseonlyknownfunctionistoproduceROS[i.e., NOX1,NOX2(akagp91phox)NOX3,NOX4,andNOX5],madeitclearthatROSarenotmerely wasteortoxicby-productsthatneedtoberemovedinordertopreventcellulardamage(Box1).
Incontrast,ROSarepartofsignalingnetworksthatincludeROStargets,suchasnitricoxide synthase(NOS)andsolubleguanylatecyclase(sGC),andROS-metabolizingenzymes,in particularthosegeneticallyregulatedbynuclearfactor(erythroid-derived2)-like2(NRF2) [3,4].The differentialexpression of theseplayers in differentsubcellular compartments [5]
suggeststhatahomogenouscell-wideredoxlevelorbalancedoesnotexist,butinfactpoints totherebeingdifferentialasynchronoushotspotsofROSsignalingwithinthecell.Thisthen makesROSquitesimilartootherclassicalsignalingmechanisms[6],includingthoseinducing post-translationalproteinmodificationssuchasphosphorylation.
ROSsignalingcontributestophysiologicalfunctionsandprocessessuchastheoxidativeburst of the innate immune response, cell proliferation and angiogenesis, vasodilation, hearing, hormonesynthesis,insulinsecretion,andinsulinsensitivity[3,7,8].DysfunctioninROSsignal- ingincludesformationofexcessiveamountsofROS,appearanceofROSatnonphysiological subcellularsitesorincelltypesthatnormallydonotformrelevantamountsofROS,orshifting fromaphysiologicaltoanonphysiologicaltypeofROS[e.g.,fromhydrogenperoxide(H2O2)to superoxide].
ThesemechanisticinsightsarenowleadingtonewtherapeuticconceptsforwhichDMisone of best understood and most suitable pathologies. Moving forward, ROS-related drug developmentwillhavetofocusonthedelicatetaskofidentifyingthemaindisease-relevant sourcesofdeleteriousROSwhileatthesametimeleavingessentialphysiologicalsourcesof ROSandtheirsignalingpathwaysintact.Oneexampleistheselectiveinhibitionofspecific NOX isoforms orthe use of NRF2 agonists that enhance the expression of endogenous antioxidantenzymesattheirphysiologicalsites,whichisqualitativelydifferenttoexogenous scavengingantioxidantsactingbroadly,inallcellsandallcellularcompartments,andthusina nonphysiological manner. As a complicating factor, during the course of disease, the enzymaticsources ofROSmaychange. Oneexample ofthis isthe triggeringofvascular dysfunctionbyNOX1,leadingtouncouplingofNOS3,whichmaybecomeasecondary,yet quantitativelymorerelevantsourceofROS[9];anotherexampleisROS-inducedROSrelease inmitochondria [10].
Box1.NADPHOxidaseFamily
UnlikeseveralROSsources(e.g.,mitochondria,xanthineoxidase,cytochromeP450enzymes,anduncoupledNOS), NADPHoxidasesaretheonlyenzymefamilyknowntoproduceROSastheirprimaryandsolefunction[5].NADPH oxidasesareenzymecomplexeswithamembranespanningcatalyticNOXsubunitinadditiontoothermembraneand cytosolicproteins.TherearesevenidentifiedmembersoftheNADPHoxidasefamily,NOX1–5,andtwodualoxidases (DUOX),DUOX1andDUOX2.Inhumans,allsevenenzymesareexpressedandeachNOXisoformhasspecifictissue expression,regulation,andtypeofROSproduced[3,5].
ThemainNOXisoformstobeconsideredinpathophysiologyofDMareNOX1andNOX5,whichproducesuperoxide, andNOX4,whichproducesH2O2.NOX3andtheDUOXshaveverylimitedrolesintheinnerearandinthesynthesisof thyroidhormone,respectively[87].NOX2isakeyenzymeoftheinnateandinflammatoryresponseanditsinhibitionor geneticdefectsareassociatedwithimmunedeficiencyandincreasedriskofinfection,inparticularinDM[13].In addition,NOX2hasbeensuggestedtobeinvolvedinanexcessiveandunlikelynumberofdiseasemodels[5],which indicatesapossiblepositivepublicationbias,asshownbyameta-analysisofNOX2studiesinstroke[88],oran epiphenomenawithouttherapeuticrelevance.
InadditiontopreventingROSsynthesisandmetabolism,ROS-induceddamagecanalsobe repaired at both the molecular and functional level. This is exemplified by recoupling of uncoupledNOS,allostericsensitizationofthenitricoxide(NO)receptor,sGC,forlowered NOlevels,andtheactivationofoxidativelydamaged,heme-freeapo-sGCbyheme-mimetics, respectively[3,5].Collectively,thesediscoverieshaverevolutionizedthefieldofROSingeneral and,asindicatedbyoneoffirstclinicalapplications,DM.
We here review, first, the physiological roles of ROS (e.g., NO, H2O2, superoxide, and peroxynitrite)ininsulinsecretionandsignaling(Box2),aswellas inphysiologyandpatho- physiologyoforgansaffectedbyfunctionalandstructuraldamageduetotype2DM(T2DM).
Importantly,the sourcesandmechanisms ofROS differin health,early andlate stagesof
Glossary
Atherosclerosis:thenarrowingof arteriesduetoplaquebuild-upon thearterialwall,achronic inflammatoryconditionoflarge-to medium-sizedarteriesandmajor causeofcardiovasculardisease.
Glomerularhyperfiltration:a conditionofabnormallyhigh glomerularfiltrationrateinthekidney, whichoccursasearlymanifestation ofdiabeticnephropathy.
Insulinresistance:astatewhen bodycellsdonotrespondsufficiently toinsulinandadrivingfactorleading toT2DM.
NADPHoxidase:amembrane- boundenzymefamilywhichconsists ofsevenmembers,NOX1–5andtwo dualoxidases(DUOX1and2).They aretheonlyknowndedicatedsource ofROSandhavenootherfunction.
Nitricoxidesynthase(NOS):a familyofenzymesthatconsistsof threeisoforms;NOS1–3,catalyzing theformationofNO,animportant cellularsignalingmolecule,fromL- arginine.
Nuclearfactor(erythroid-derived 2)-like2(NRF2):atranscription factorthatregulatestheexpression ofcytoprotectiveandantioxidant proteins.
Oxidativestress:anoutdatedview ofROSasprimarilytoxicmetabolic wasteproductsthatneedtobe controlledbyantioxidantenzymesor antioxidantdrugs.
Reactiveoxygenspecies(ROS):a collectivetermforcompoundssuch assuperoxide,hydrogenperoxide, hydroxylradical,peroxynitrite,and nitricoxide,althoughreactivityvaries greatlybetweenindividualmembers.
Redoxbalancehypothesis:an outdatedhypothesisthatwas frequentlyusedtodescribea supposedlyneutralsteady-statelevel ofROS(i.e.,balancebetweenROS formationandcellularantioxidant defense).
Renin-angiotensinsystem(RAS):
ahormonesystemwhichcontrols bloodpressureandsodium homeostasis.Itseffectsare coordinatedthroughintegrated actionsinthekidney,cardiovascular system,andthecentralnervous system.
Solubleguanylatecyclase(sGC):
aheme-containingenzymethatacts asareceptorforNOtoproducethe Box2.InsulinandROS
PancreaticbCells
Thesecretionofinsulinfrompancreaticbcellsismainlyregulatedbyplasmaglucoselevels.Glucoseistakenupviathe glucosetransporter2(GLUT2)andthereafteroxidizedtoproduceATP.ThisleadstotheclosureofKATPchannels, depolarizationoftheplasmamembrane,openingofvoltage-dependentCa2+channels,andasubsequentincreasein intracellularCa2+,whichenablesexocytoticinsulinrelease[89].TwoROSspecies,NOandH2O2,increaseintracellular Ca2+andtherebyfacilitateinsulinrelease.LowlevelsofNOfromNOS1dosobystimulatingcGMPformationthrough sGC,andcGMPinturnactivatescGMP-dependentproteinkinase(PKG),whichinhibitsKATPchannels[90];mitochon- drialglucosemetabolismleakssmallamountsofsuperoxide,which,upondismutationtoH2O2,stimulatesryanodine receptorstoalsoincreaseintracellularCa2+[7,89].Importantly,theuseofROSbypancreaticbcellscomesatarisk.
Comparedwithmostothercells,pancreaticbcellshavesomeofthelowestexpressionandactivitylevelsofROS- metabolizing(antioxidant)enzymes[7],makingthemmorevulnerablethanothercellstothepotentialcytotoxiceffectsof ROS,suchasDNAandproteindamage.
InearlystagesofT2DM,productionofNOandROSbeginstoexceedtheantioxidantresistanceofpancreaticbcells.
Peripheralinsulin resistance(seebelow) and highglucoseand fattyacidlevelstriggerpancreatic bcells to compensatebyreleasingmoreandmoreinsulininamitochondrialROS-dependentmanner[91].Inlaterstages,this mechanismisexhaustedassuperoxideanioninducesaleakofprotonsacrossthemitochondrialinnermembrane, decreasesthemitochondrialmembranepotentialandATPproduction,andeventuallyleadstoalowerinsulinsecretion [92].Moreover,upregulationoftherenin-angiotensinsystem (RAS)andinflammatorycytokines furtherincrease superoxideoverproductionfromadditionalsources(Figure1)(i.e.,NOX1andNOX2)[87].Theadditionalinduction ofNOS2,whichproducesNO,toxifiessuperoxidetoyieldperoxynitritewhichnotonlyfurtherreducesinsulinsecretion butalsoinducespancreaticbcelldeath[93].
Insulin-SensitiveTissues
Inperipheralinsulin-sensitivetissuessuchasskeletalmuscle,fatcells,andliver,insulincontrolstheswitchfromlipolysis/
fattyoxidationduringfastingtolipidstorage/glucoseoxidationfollowingfeeding[94].Bindingofinsulintotheinsulin receptor(IR)phosphorylatessubstrateproteins,IRS1andIRS2,activatingphosphatidylinositol3-kinase(PI3K)-Akt (proteinkinaseB)signaling,whichleadstothetranslocationandactivationofglucosetransporters(mainlyGLUT4in musclesandfatcells)andsubsequentglucoseuptake[94,95].
ROScomeintoplayininsulinsignalingthroughactivationofPI3KandalternativeproteinkinaseC(PKC)activationto increaseNOX4activity,formingH2O2[95].H2O2augmentsinsulin-IR-PI3Ksignalingtwofold,byinhibitingprotein tyrosinephosphatase1B(PTP1B)andthephosphataseandtensinhomologue,PTEN,whichdephosphorylatesIRand downregulatesPI3Ksignaling[94,95],andbyactivatingMAPkinasephosphatase-1,whichdephosphorylatesIRS1 [96].
FurtherincreasedROSproductionisassociatedwithperipheralinsulinresistance,amainfeatureofT2DM[97].Inearly stagesofT2DM,NOX4causesfatandlivercellapoptosis,fibrosis,andinflammation[97,98];NOX2decreasesskeletal muscleinsulin-inducedAktphosphorylation,GLUT4expressionandtranslocation,andtherebyglucoseuptake[99].In laterstages,mitochondrialROSformationisinducedinskeletalmuscles,fatcells,andliver,activatingserinekinases andfurtherimpedinginsulinsignaling[97,98].Inskeletalmuscles,upregulationoftheRASsystemfurtheractivates NOX2toaggravateinsulinresistance[99].
314 TrendsinEndocrinology&Metabolism,May2019,Vol.30,No.5
secondmessengercGMP,which activatesproteinkinasesandcauses vasodilationorregulates
phosphodiesterases.
Tubuloglomerularfeedback:an adaptivemechanismthatlinksthe rateofglomerularfiltrationtothe concentrationofsaltinthedistal tubules.
UncoupledNOS:whenNOS enzymaticactivityisuncoupledto producesuperoxideinsteadofNO.
Thisstateusuallyoccursunder pathologicalconditionsinducedby ROS.
disease,andwithrespecttodifferentcellsandorgans(Figure1,KeyFigure),sometimeseven qualitatively.ThesedifferencesneedtobecarefullydissectedanddefinedforT2DMandits characteristic complications as well as other conditions in order to allow future precision antidiabeticinterventions.WithrespecttoROS,therelevanttherapeutictargets(seeabove) form acausal signaling network andin T2DM will bebest targeted by mechanism-based networkpharmacology,wheremultipledrugsarecombinedtosynergisticallycorrectapath- ologicalintoanear-physiologicalstate[11].Thesedrugsincludesomenewcompoundclasses thatactdirectlyonROSsourcesandtargetsbutalsosomeregistereddrugs,whichact,atleast inpart,throughindirectlymodulatingtheROSnetwork.Finally,mechanism-baseddiagnostics mayenrichthesenewtherapeuticapproacheswithrespecttoROSandDMinordertostratify patientsforpersonalizedandprecisiontherapy.
ROSinDiabeticEnd-Organ Damage/Injury
Themost patient-relevant end-organ complications in DM include chronic kidney disease (CKD)(27.8% ofpatients withDM), retinopathies (18.9%), heart attack(9.8%),andstroke (6.6%)[12].Inallofthese,ROShavebeensuggestedtoplayacausalrole[13–15].Collectively, theserepresentmajor causes ofdisability anddeath indiabeticsand areonly moderately preventedbymostglucose-loweringantidiabeticdrugs,inparticulardiabetickidneydisease [16]. AlthoughROS play detrimental rolesin thesecomplications, certain ROS forms fulfil importantphysiologicalfunctionsinseveralorgans,forexample,NO,whichisproducedby constitutiveNOSenzymes(i.e.,NOS1andNOS3)andH2O2,whichisproducedbyNOX4.
Someexamplesofthesearediscussedinthefollowingsections.
Indeed,inT2DM,severaleventssuchashyperglycemia,dyslipidemia,advancedglycationend- products(AGEs),andupregulationoftherenin-angiotensinsystem(RAS)contributetodetri- mentalROSproduction.Despitethefactthatmanyenzymesare,inprinciple,capableofforming ROS,NOXsappeartooftenbe theprimarysourceanddiseasetrigger[14].Importantly,they representtheonlyenzymefamilywithnootherknownfunctionthantoproduceROS[5].Allother ROSsourceshaveotherprimaryfunctionsandROSproductionisabiochemical‘accident’often requiringaprior(oftenROS-induced)damageoruncouplingbeforeROSformationisinitiated.
Therefore, we will focushereonthe roleof mainNOX isoforms(Box1) in diabetic end-organdamage.
TherolesplayedbyotherROSsources,suchasmitochondriaandxanthineoxidase(XO),andthe ROS-toxifyingenzyme,myeloperoxidase(MPO)aswellasNRF2arebrieflydiscussedinBox3.
ROSinBloodVessels
In anormalblood vesselwall, NO produced byNOS3 activates sGC,resultingin cGMP- dependent inhibition of smooth muscle contraction. In human vascular cells, NO shows atheroprotection by inhibiting leukocyte adhesion in endothelial cells and proliferation of smoothmusclecells[17,18].NOalsodisplaysantithromboticpropertiesinhumanendothelial cellsbyinhibitingplateletadhesionandaggregation[18].H2O2producedbyendothelialNOX4 enhances vasodilation inmice [8],both ina cGMP-dependent manner, by increasingthe expressionandactivityofNOS3 [19,20],andinacGMP-independentmanner,byoxidative activationofproteinkinaseGI(PKGIa)[21].Moreover,H2O2isakeysignalinangiogenesisin bothhumanandanimalvascularcells[22,23].
VascularDisease/Atherosclerosis
Inbloodvessels,ROShavebeensuggestedtocausehypertension,atherosclerosis,anda prothromboticstage,eitherdirectlyorbyinterferingwithprotectiveNO[13,24].Inlarge-and medium-sizedarteries,ROS-inducedatherosclerosisisthoughttobefurtheracceleratedby DM [13].Surprisingly, however, this doesnotaccount forall ROS; NOX4-derivedH2O2is
KeyFigure
Differential Roles of ROS in Physiology and DM
Glucose
GLUT2 GLUT1,2,4
Glucose Insulin Glucose Insulin
Glucose Glucose Insulin Glucose Insulin
FFA FFA FFA
GLUT1
GLUT2 GLUT 1 2GLUT 1,2,4GLUT1 2 4T 4 GLUT1 IR
IR IR IR IR
IR
ATP
Insulin release
NOX4
NOX4 NOX4
NOX4 NOX4
NOS3
NOX5
PI3K PI3K
PI3K PI3K
NOS1,3
NO NO
Urine excr
Angiogenesis
Differe
Hyperfiltra n
Hypofiltr sGC
sGC
cGMP cGMP
Dilata n
NOX4 NOS1,3
NOS
NO
NO
sGC
sGC
cGMP
cGMP Dilata n –
Glomerular hypertrophy
?
PhysiologyEarly diabetes
Glucose Glucose Glucose nsulin
FFA
GLUT2 GLUT1,2,4 IR GLUT1 IR
Ang II Cytokines FFA Ang II
ATP
Insulin release
NOX4,5 NOX4 PI3K
PI3K NOX1
ATP –
NO NOX1 NOS2
PKC
ROS uc-NOS Atherosclerosis
Insulin release Fibrosis,
Apo-sGC
β cell Kidney Endothelium
smooth muscle
NOX4 NOX1,5
uc-NOS
β cell
death Fibrosis, prolifer
–
Lategan damage
–
Apo-sGC Ang II
(A)
(B)
(C)
GLUT1
FFA
–
GLUT1 GLUT1
(Seefigurelegendonthebottomofthenextpage.) 316 TrendsinEndocrinology&Metabolism,May2019,Vol.30,No.5
antiatheroscleroticbyreducing fibrosisand proliferationof smoothmuscle cells [25].Con- versely,adifferenttypeandsourceofROS(i.e.,NOX1-[13]andpossiblyNOX5-derived[26]
superoxide),maybeproatheroscleroticinDM.TheseexamplesshowthecomplexityofROS pathobiologywithdifferentsources/typesofROShavingqualitativelyopposingeffects,making precisetargetingof themost disease-relevant isoformpertinentfor any chronictherapy in T2DM.Nonspecific,chronicNOXinhibitionmayinterferewithangiogenesis,collateralforma- tion,capillarization,andmayalsobeproatheroscleroticandimmunocompromising.
Figure1.Panelsin(A)depictphysiologiceffectsofROS;(B),pathologicalchangesinearlyDM;(C),late-stageDMend- organdamage.Ineachrow,theblackboxesontheleftrepresentpancreaticbcells;redinthemiddle,renalcells;green andblueontheright,endothelialandvascularsmoothmusclecells,respectively.Blackarrowsandredblocksrepresent stimulatoryand inhibitory effects,respectively. ROS mediate bothphysiologicaland pathophysiologicalsignaling.
Mechanismscanbeenhanced(boldblacktextandarrows)ordiminished(small/transparenttextandarrows)leading topathologicalchangesthatareimplicatedinearlyandlatestagesofDM.Abbreviations:AngII,angiotensinII;DM, diabetesmellitus;FFA,freefattyacids;GLUT,glucosetransporter;H2O2,hydrogenperoxide;IR,insulinreceptor;NO, nitricoxide;NOS,nitricoxidesynthase;NOX,NADPHoxidase;O2,superoxide;PI3K,phosphatidylinositide3-kinase;
PKC,proteinkinaseC;ROS,reactiveoxygenspecies;sGC,solubleguanylatecyclase.
Box3.ROSSourcesandMetabolism Mitochondria,XO,andMPO
InadditiontoNOXsanduncoupledNOS,mitochondriaandXOasalternativeROSsources,andMPOasanimportant ROS-toxifyingenzymemayalsocontributetothepathogenesisofdiabeticend-organdamage.XOproducesROSsuch assuperoxideandH2O2duringoxidativeconversionofxanthinetouricacid[5].Mitochondriaproducesuperoxide,by oneelectronreductionofO2,whichisthendismutatedtoH2O2bySOD[100].MPOisahemeperoxidaseexpressedin neutrophilsandmonocytesandisimportantforinnateimmunesystem.MPOconvertsNOX-orXO-derivedH2O2in conjunctionwithhalidesandnitritetomorereactivespeciessuchashypochlorousacid[5]and,inconjunctionwith nitrite,representsanalternativesourceforperoxynitrite,besidesthecanonicalNOandsuperoxideinteraction[101].
Importantly,elevatedXOactivityandMPOlevelscorrelatewiththedevelopmentofT2DManddiabeticend-organ damage[102,103](e.g.,inkidneyfibrosisandproteinuria[104],neuropathy[105],andatherosclerosis[103]).Mito- chondrialROSalsohavebeensuggestedascausativefactorsofinsulinresistanceandimplicatedininitiationand progressionofdiabeticcomplications[14].Therefore,targetingmitochondria,XO,andMPOmaybeofadditionalbenefit inDMand,indeed,mitochondrial-targetedantioxidants,XO,andMPOinhibitorsareinclinicaldevelopment(seebelow) (Table1).
NRF2
BesidesROS-formingand-metabolizingenzymes,thereappearstoalsobeatherapeuticoptionviaendogenousROS- eliminating(antioxidant)enzymes,inparticularthosegeneticallyregulatedbythetranscriptionfactorNRF2.Invascu- lature,NRF2isstimulatedinresponsetosteadylaminarshearstressandROStopromoteatheroprotection.Con- versely,oscillatoryshear stress andendothelial dysfunction are associatedwith decreased NRF2 activityand atherogenesis[106].Inthediabeticmilieu,NRF2anditstargetgenesincreaseinresponsetohyperglycemiaand highfat,suggestingadaptiveactivationagainstincreasedROS[107].However,chronichyperglycemiaandprolonged ROSproductioninhibitNRF2activation,whichisassociatedwithendothelialcelldeathandfoamcellformation[108].In thediabetickidney,highglucoseandincreasedROSproductionareaccompaniedbyNRF2upregulation,whichinitially protectsagainstinjuryviainhibitionofTGF-bandaccumulationofextracellularmatrix[109].Sustainedhyperglycemia andformationofAGEs,however,downregulateNRF2anddecreaseitsactivityinthekidney,whichisassociatedwith increasedfibrosisandrenaldysfunction[110].Moreover,inCKDpatients,adecreaseintheNRF2signatureof peripheralbloodmononuclearcellscorrelateswithincreasedNF-kBrelatedproinflammatoryresponses[111].Retinal NRF2activationisincreasedinDM,whichisthoughttoprotectagainstROSoverproductionandinflammatorycytokines aswellasblood–retinabarrierdysfunction,however,theDNAbindingactivityofNRF2isdecreased[112].Similarly,in diabeticnerves,acuteandchronichyperglycemiaareassociatedwithincreasedanddecreasedlevelsofNRF2, respectively[113].Thus,inadditiontoNOX,XO,MPO,NOS,andsGC,NRF2alsohastobeconsideredforthe treatmentorpreventionofdiabeticend-organdamageandallofthesetargetsmayformacausalinter-relatedsignaling networkformechanism-basednetworkpharmacology(Figure2).
Superoxide appears to be the most disease-relevant type of ROS. It can decrease NO biophaselevelsbydirectchemicalscavenging,leadingtointermediateperoxynitrite,protein tyrosinenitration,reducingendothelialinsulinreceptorexpression,andinhibitingphosphati- dylinositol3-kinase(PI3K)-Akt-NOS3signalingintheendothelium[24,27](Figure1).Inaddition, superoxideuncouplesNOS3,whichdecreasesNOproductionandsimultaneouslyincreases superoxideproductionfromuncoupledNOS3[9](i.e.,anexampleofROS-inducedROS).This hasalsobeencoinedasthe‘kindling-bonfire-radical’hypothesis[28](Figure1),becausethe totalamountsofROSformedfromuncoupledNOScaneventuallyexceedthoseoftheinitial triggerenzyme,NOX.Finally,superoxideand/orperoxynitritecandamagetheNOreceptor, sGC [29,30], leading to a collectively threefold interruption of NO-cGMP signaling by:
(i)scavengingofNO,(ii)uncouplingNOS3,and(iii)damagingtheNOreceptorsGC.
ROSintheKidney
In normalkidney, ROS regulate urine excretionand blood pressure, and H2O2stimulates prorenin-inducedsodiumreabsorptioninmousekidney[31]andvasodilatesintrarenalarteries from human and rat through indirect NO-dependent mechanisms [19,32]. NO promotes natriuresisin animals byinhibiting tubular sodiumreabsorption [33],promotes diuresisby increasing both total and regional renal blood flow through functional antagonism of the vasoconstrictortoneinducedbyrenalsympatheticneurons[33,34],andbluntsthetubulo- glomerularfeedbackresponsebyafferentarteriolardilatation[35].
DiabeticKidneyDisease
Besides ROS, multiple inflammatory, fibrotic, and apoptotic signaling pathways are also implicatedinthedifferentstagesofdiabetickidneydisease.ROS,however,appeartointegrate theseandtherebyplay acrucial roleintheinitiationandprogressionofdiabeticend-organ damageandthusrepresentanidealtarget[14].Forthis,diabeticnephropathyrepresentsthe clinicallymostadvancedtherapeuticapproach(Table1),wheredisease-relevant(notprotec- tive!)H2O2isproducedfrom NOX4,andsuperoxide,from NOX5.Despitebeingchemically distinct,bothROSseemtotriggerkidneyfibrosis,renalhypertrophy,andalbuminuria[36,37].
H2O2alsoincreasestheexpressionofvascularendothelialgrowthfactor(VEGF)andprofibrotic markers[36]andinducesglomerularhyperfiltration,possiblybyincreasingintrarenalNOS expression[38].Whilethediseaseprogresses,severalotherrenalmediators,suchasangio- tensin II, AGEs, transforming growth factor (TGF)-b, and protein kinase C (PKC), further increasethe activityof NOXenzymes,resultinginafurther aggravatedrenal damage[14].
ImpairedcGMPsignalingbyROSuncouplingofNOS3andoxidizingsGC(Figure1)mayfurther aggravatetubulointerstitialdamageandfibrosis[39].
ROSinOtherOrgans
Intheretina,NOisnecessaryforvisualfunctionbyincreasingretinalbloodflow,tomaintain adequatenourishmentandmeetthehighmetabolicdemandoftheretina.Inanimalretina,NO modulatessynaptictransmissionfromphotoreceptorsandactivatessGCtoproducecGMP, whichisakeyintermediateinthevisualtransductioncascade(reviewedin[40]).H2O2mayalso actas anintracellularmessengerinthe humanretina.Itsphysiological roles,however,are unclear[41].
Inthebrainofanimals,NOfunctionsasaneurotransmitterandneuromodulator,regulating synapticplasticity,which isinvolvedincognitive functionssuch as memoryformation and mediatesneurovascularcoupling[42].H2O2contributestoneurogenesisanddifferentiation, andneuronalplasticity[43].Intheperipheralnervoussystemofanimals,NOalsoservesasa neurotransmitter of so-called nonadrenergic-noncholinergic nerves, which induce smooth
318 TrendsinEndocrinology&Metabolism,May2019,Vol.30,No.5
Table1.Mechanism-BasedTherapeuticsandTheirClinicalStatus
Compound Mechanismofaction Pathology Clinicalstatus
GKT137831 NOX1,NOX4,NOX5inhibitor
T2DMnephropathy SafeinPhaseIclinicaltrial
Failedtoreducealbuminuriainashort-term PhaseIItrial(NCT02010242)
T1DMnephropathy Clinicaltrialongoing(U1111-1187-2609) Primarybiliarycirrhosis PhaseIIclinicaltrialongoing(NCT03226067)
Ronopterin(VAS203) NOS2inhibitor
Traumaticbraininjury PhaseIIIclinicaltrial(NCT02794168) Renalfunctionin
healthyvolunteers
SafeinPhaseIclinicaltrial(NCT02992236)
Allopurinol XOinhibitor T1DMnephropathy PhaseIV(NCT02829177)
AZD3241 MPOinhibitor Multiplesystematrophy PhaseII(NCT02388295)
AZD4831 MPOinhibitor Heartfailure PhaseII(NCT03756285)
MitoQ Mitochondrial-targetedantioxidant Chronickidneydisease Declaredasdietarysupplementwithout indicationorapplicationinPhaseIV (NCT02364648)
SkQ1 Mitochondrial-targetedantioxidant Dry-eyesyndrome PhaseIII(NCT03764735)
Folicacid NOSrecouplingagent T2DMnephropathy Failedtoimproverenalendothelialfunctionor
reducealbuminuriainaPhaseIIIclinicaltrial (NCT00136188)
L-citrulline NOSrecouplingagent
VasculardysfunctioninT2DM Clinicaltrialongoing(NCT03358264) Peripheralarterydisease PhaseIIIclinicaltrialongoing(NCT02521220) Sicklecelldisease PhaseItrialongoing(NCT02697240)
Riociguat sGCstimulator
Pulmonaryhypertension Intheclinic
Scleroderma PhaseIItrialongoing(NCT02915835)
Sicklecelldisease PhaseIItrialongoing(NCT02633397)
Vericiguat sGCstimulator Heartfailure PhaseIIIclinicaltrial(NCT02861534)
Olinciguat(IW-1701) sGCstimulator
TypeIorIIachalasia PhaseIIclinicaltrialongoing(NCT02931565) Sicklecelldisease PhaseIIclinicaltrialongoing(NCT03285178)
Praliciguat(IW-1973) sGCstimulator
Heartfailurewithpreserved ejectionfraction
PhaseIItrialongoing(NCT03254485)
T2DMandhypertension PhaseIItrialcompleted(NCT03091920) Nelociguat(BAY60-4552) sGCstimulator Chronicheartfailure PhaseIcompleted(NCT00565565)
Ataciguat sGCactivator Aorticvalvecalcification PhaseIIclinicaltrial(NCT02481258)
Bardoxolonemethyl NRF2activator
T2DMandstage4CKD IncreasedestimatedGFR,butdidnotimprove proteinuriaandwasassociatedwith cardiovasculartoxicityinPhaseIIIclinicaltrial (NCT01351675).
T2DMandCKD IncreasedmeasuredGFRwithnoheartfailure eventsduetofluidretentioninPhaseIIclinical trial(NCT02316821)
CKDassociatedwithT1DM PhaseIIclinicaltrial(NCT03366337)
CXA-10 NRF2activator
Pulmonaryarterialhypertension PhaseIIclinicaltrialongoing(NCT03449524) Primaryfocalsegmentalglomerulosclerosis PhaseIIclinicaltrialongoing(NCT03422510)
muscle relaxation [42].H2O2 stimulates axon growth and nerve regeneration [43,44]. The humancellculturedata[17,18,23,32,41]aswellastheresultsinanimalsdescribedabove,if translatable to humans, would suggest that ROS contribute significantly and in multiple mannerstohumanphysiology.
OtherEnd-OrganComplications
Indiabeticretinopathy,H2O2productionisincreasedfromanupregulatedNOX4,resultingin increasedVEGFexpressionandblood–retinabarrierdamage[45].Moreover,NOX1isacti- vatedtoproducesuperoxide,whichpromotescelldeath[46].Inchronicdiabeticneuropathy, ROSfromayetundefinedisoformofNADPHoxidaseareassociatedwithneuronalapoptosis [47].Inacuteischemicstroke[48],NOX4isthemainsourceofdeleteriousROSandinduces blood–brain barrier (BBB) damage and neuronal cell death [15,49]. In the setting of DM, additionalisoforms(e.g.,NOX1andNOX5)maycomeintoplay.
MechanismandNetwork-BasedRedoxTherapies
Consideringthefailureofclassicantioxidantsinclinicaltrialsfocusingondiabeticcomplications [50,51],newtherapeuticapproaches(i.e.,mechanism-basedredoxtherapies)arenowstate- of-the-artfor futuretrials. These approachesincludetargeting specificROS sources using pharmacological inhibitors, repairing ROS-induced damage, or upregulating endogenous antioxidant enzymes (Figure 2). The clinical status of these therapies is listed inTable 1.
Importantly,targetingtheROSsignalingnetworkatmultiplesitesinasynergisticcombinatorial mannerenablesashiftfrom singletargetapproachestonetworkpharmacology,facilitating lowertherapeuticdoseswithfewersideeffects.
DrugsTargetingROSSources NOXInhibitors
SeveralNOX inhibitorsexist [3].However,noneof themare isoformspecific.TheNOX1/4 inhibitors,GKT137831andGKT136901,haveshownbeneficialeffectsinseveralpreclinical studies addressing diabetic complications [13,36]. GKT137831 is currently the only NOX inhibitorinclinicaltrialsfocusingondiabeticnephropathy.ItwassafeinaPhaseItrial,however failedtoreducealbuminuriainashort-termPhaseIItrialinpatientswithT2DMandadvanced nephropathy on maximal RAS blockade. Yet, there were positive signals on reduction of systemicinflammatorymarkers.Anothertrialiscurrentlyinvestigatingtheantialbuminuriceffect ofGKT137831intype1diabeticnephropathy.Thiscompoundisalsobeingtestedinaclinical trialforprimarybiliarycirrhosis(Table1).
InadditiontoGKT137831,otherNOXinhibitorsshowedpromisingresultsinpreclinicalstudies well,thepan-NOXinhibitor,APX-115,protectsagainstnephropathyinamouseT2DMmodel andwassuperiortoGKT137831inpreservingrenalfunction[52].VAS2870,anotherpan-NOX inhibitor,reducesaorticcontractilityindiabeticratsbyimprovingendothelialfunction[53]and, instroke,stabilizedtheBBB,providedneuroprotection,improvedneurologicalscoresinmice [15],andinhibitedreperfusion-inducedhemorrhagictransformationinhyperglycemicrats[54].
Clearly,isoform-specificNOXinhibitorstogetherwithfurthertestinginclinicaltrialsfocusingon diabeticcomplicationsareneeded.
NOSInhibitors
Despiteitsmanybeneficialactions,NOmayalsohavedetrimentaleffectswhenoverproduced.
Afterischemicstroke,forexample,overproductionofNObyNOS1inducescelldeathandBBB hyperpermeability and in preclinical studies both nonselective NOS inhibitors or selective NOS1/2 inhibitors improved post-stroke outcome [55].Thus far, however, no clinical trial
320 TrendsinEndocrinology&Metabolism,May2019,Vol.30,No.5