NANOPARTICLE-ASSISTED IMMUNOASSAYS FOR POINT-OF … · 2017-02-22 · reductions in the assay...

1

Heidi Hyytiä

TURUN YLIOPISTON JULKAISUJA – ANNALES UNIVERSITATIS TURKUENSISSarja - ser. AI osa - tom. 510 | Astronomica - Chemica - Physica - Mathematica | Turku 2015

NANOPARTICLE-ASSISTED IMMUNOASSAYS FOR

POINT-OF-CARE TESTING– With Specific Interest in Minimally

Interference-Prone Assays for Cardiac Troponin I

2

Supervised by

Professor Kim Pettersson, PhDDepartment of Biochemistry Molecular Biotechnology and DiagnosticsUniversity of TurkuTurku, Finland

Eeva-Christine Brockmann, PhDDepartment of BiochemistryMolecular Biotechnology and DiagnosticsUniversity of TurkuTurku, Finland

University of Turku

Faculty of Mathematics and Natural SciencesDepartment of BiochemistryMolecular Biotechnology and DiagnosticsDoctoral Programme in Molecular Life Sciences

Reviewed by

Susann Eriksson, PhDDHR Finland Oy Innotrac DiagnosticsTurku, Finland

Petri Saviranta, PhDVTT Technical Research Centre of FinlandTurku, Finland

Opponent

Senior Consultant Kjell Nustad, M.D., Ph.D.Department of Medical BiochemistryOslo University Hospital, the Norwegian Radium HospitalOslo, Norway

The originality of this thesis has been checked in accordance with the University of Turku quality assurance system using the Turnitin OriginalityCheck service.

ISBN 978-951-29-6065-1 (PRINT)ISBN 978-951-29-6066-8 (PDF)ISSN 0082-7002Painosalama Oy - Turku, Finland 2015

3

Thewaytogetstartedistoquittalkingandbegindoing.

‐WaltDisney‐

Contents

4

CONTENTS

CONTENTS.............................................................................................................................4LISTOFORIGINALPUBLICATIONS...............................................................................6ABBREVIATIONS.................................................................................................................7ABSTRACT.............................................................................................................................9TIIVISTELMÄ......................................................................................................................101 INTRODUCTION...........................................................................................................112 LITERATUREREVIEW................................................................................................132.1CardiactroponinI...............................................................................................13

2.1.1 ImmunoassaysforcardiactroponinI...........................................................132.1.2 Point‐of‐caretesting.............................................................................................14

2.2 Lanthanide‐dopedpolystyrenenanoparticles.........................................152.2.1 Lanthanidechelatetechnology........................................................................152.2.2 Productionoflanthanide‐dopedpolystyrenenanoparticles.............152.2.3 Functionalizationofpolystyrenenanoparticles......................................162.2.4 Immunometricassaysutilizingpolystyrenenanoparticles................17

2.3 Immunoassayinterferences............................................................................172.3.1 Interferencemechanismsofcirculatinghumanantibodies...............182.3.2 Heterophilic‐typeinterference........................................................................19

2.3.2.1Humananti‐animal,anti‐chimericandanti‐humanantibodies...................................................................................................20

2.3.2.2Heterophilicantibodies.......................................................................212.3.2.3Rheumatoidfactors...............................................................................222.3.2.4Incidenceofheterophilic‐typeinterference...............................22

2.3.3 Autoantibodiesagainsttheanalyte...............................................................232.3.4 Otherproteins.........................................................................................................242.3.5 Hemolysis,icterusandlipemia........................................................................242.3.6 Cross‐reactivity.......................................................................................................252.3.7 Pharmaceuticalpreparations...........................................................................252.3.8 Sampleprocessing.................................................................................................252.3.9 High‐dosehookeffect..........................................................................................26

2.4Heterophilic‐typeinterference:preventionandelimination..............272.4.1 Immunosuppressanttherapy...........................................................................272.4.2 Immunoassaydesign............................................................................................28

2.4.2.1Enzymaticallycleavedfragments....................................................282.4.2.2Recombinantantibodiesandantibodyfragments...................29

2.4.3 Physicalandchemicaltechniques..................................................................312.4.3.1Precipitation.............................................................................................322.4.3.2Affinityextraction...................................................................................322.4.3.3Otherapproaches...................................................................................32

2.4.4 Blockingagents.......................................................................................................333 AIMSOFTHESTUDY..................................................................................................36

Contents

5

4 SUMMARYOFMATERIALSANDMETHODS........................................................374.1 Samples..................................................................................................................374.2Reagents.................................................................................................................38

4.2.1 Antibodies.................................................................................................................384.2.2 Calibrators................................................................................................................39

4.2.2.1D‐dimer.......................................................................................................394.2.2.2TroponinI..................................................................................................39

4.2.3 Otherreagentsandassaybuffers...................................................................394.3Preparationofassayreagents........................................................................40

4.3.1 Carboxyl‐modifiednanoparticles...................................................................404.3.2 Labelingofsolid‐phaseantibodies................................................................41

4.4 Immunoassays.....................................................................................................414.4.1 Immunoassaywithpreincubationstep(I).................................................414.4.2 ImmunoassaysforcardiactroponinI(II–IV)............................................43

4.5Assayevaluations................................................................................................444.5.1 Methodologicalevaluation(I,III,IV)............................................................444.5.2 Methodcomparisons(I,III)..............................................................................454.5.3 Assayinterferences(II–IV)................................................................................454.5.4 Statisticalanalyses(I–IV)...................................................................................45

5 SUMMARYOFRESULTS.............................................................................................465.1Nanoparticlesasuniversallabels(I)............................................................46

5.1.1 Adjustmentofassaydynamicrange..............................................................465.1.2 Performanceofa15‐minuteD‐dimerassay..............................................48

5.2AnalyticalperformanceofimmunoassaysforcardiactroponinI......485.2.1 Theeffectofsolid‐phaseantibody9707molecularform(II)............485.2.2 Chimericrecombinantantibodyfragments(III).....................................495.2.3 DifferentcommercialcTnI‐specificantibodies(IV)...............................51

5.3CardiactroponinIimmunoassayspecificity.............................................525.3.1 Theeffectofsolid‐phaseantibody9707molecularform(II)............525.3.2 Chimericrecombinantantibodyfragments(III).....................................535.3.3 DifferentcommercialcTnI‐specificantibodies(IV)...............................55

6 DISCUSSION...................................................................................................................586.1 Immunoassaysandtheirperformance.......................................................58

6.1.1 Nanoparticlesaslabels(I).................................................................................586.1.2 ImmunoassayversionsforcardiactroponinI(II‒IV)...........................59

6.2Matrix‐relatedinterferences..........................................................................617 CONCLUSIONS...............................................................................................................65ACKNOWLEDGEMENTS...................................................................................................67REFERENCES.......................................................................................................................69ORIGINALPUBLICATIONS..............................................................................................85

Listoforiginalpublications

6

LISTOFORIGINALPUBLICATIONS

Thethesisisbasedonthefollowingpublications,referredtointhetextbytheirRomannumerals(I‒IV):

I Heidi Hyytiä, Noora Ristiniemi, Päivi Laitinen, Timo Lövgren and KimPettersson (2014). Extension of dynamic range of sensitive nanoparticle‐basedimmunoassays.AnalBiochem446:82–86.

II HeidiHyytiä,Marja‐LeenaJärvenpää,NooraRistiniemi,TimoLövgren,andKim Pettersson (2013). A comparison of capture antibody fragments incardiactroponinIimmunoassay.ClinBiochem46:963–968.

III HeidiHyytiä,TainaHeikkilä,Eeva‐ChristineBrockmann,HennaKekki,PirjoHedberg, Tarja Puolakanaho, Timo Lövgren and Kim Pettersson (2015).Chimeric recombinant antibody fragments in cardiac troponin Iimmunoassay. Clin Biochem Published online August 8. doi:10.1016/j.clinbiochem.2014.06.080.

IV Heidi Hyytiä, Taina Heikkilä Pirjo Hedberg, Tarja Puolakanaho and KimPettersson (2015). Skeletal troponin I cross‐reactivity in different cardiactroponin I assay versions. Clin Biochem Published online January 9. doi:10.1016/j.clinbiochem.2014.12.028.

The original publications have been reproduced with the permission of thecopyrightholder.

Abbreviations

7

ABBREVIATIONS

ACS acutecoronarysyndromeBGG bovinegammaglobulinBSA bovineserumalbuminCA cancerantigenCEA carcinoembryonicantigencFab chimericfragmentantigenbindingCDR complementaritydeterminingregionCI confidenceintervalCH constantheavydomainofantibodyCK‐MB creatinekinasemuscle‐brainfractionCL constantlightdomainofantibodyCRP C‐reactiveproteincTn cardiactroponincTnI cardiactroponinIcTnT cardiactroponinTCV coefficientofvariationED emergencydepartmentEDC N‐(3‐dimethylaminopropyl)‐N’‐ethylcarbodiimideFab fragmentantigenbindingFc fragmentcrystallizableFDP fibrindegradationproductFEU fibrinogenequivalentunitsFSH follicle‐stimulatinghormoneHAAA humananti‐animalantibodyHACA humananti‐chimericantibodyHAMA humananti‐mouseantibodyHBR heterophilicblockingreagentHBT heterophilicblockingtubehCG humanchorionicgonadotropinIgG immunoglobulinclassGIgM immunoglobulinclassMIIR immunoglobulininhibitingreagentIU internationalunitLoB limitofblankLoD limitofdetectionLoQ limitofquantitationMab monoclonalantibodyMES 2‐(N‐morpholino)ethanesulfonicacidMI myocardialinfarctionmIgG murine/mouseimmunoglobulinGNHS N‐hydroxysuccinimidePEG polyethyleneglycolPOCT point‐of‐caretesting

Abbreviations

8

RF rheumatoidfactorscFv single‐chainvariablefragmentskTnI skeletaltroponinIskTnT skeletaltroponinTsulfo‐NHS N‐hydroxysulfosuccinimideTAT turn‐aroundtimeTnC troponinCTnI troponinITnT troponinTTSH thyroid‐stimulatinghormoneVH variableheavydomainofantibodyVL variablelightdomainofantibodyVTE venousthromboembolism

Abstract

9

ABSTRACT

Cardiac troponins (cTn) I and T are the current golden standard biochemicalmarkers in thediagnosisandriskstratificationofpatientswithsuspectedacutecoronarysyndrome.Duringthepastfewyears,novelassayscapableofdetectingcTn‐concentrations in >50% of apparently healthy individuals have becomereadily available. With the emerging of these high sensitivity cTn assays,reductions in the assay specificity have caused elevations in themeasured cTnlevelsthatdonotcorrelatewiththeclinicalpictureofthepatient.Theincreasedassay sensitivity may reveal that various analytical interference mechanismsexist.

Thisdoctoral thesis focusedondevelopingnanoparticle‐assisted immunometricassaysthatcouldpossiblybeappliedtoanautomatedpoint‐of‐caresystem.Themain objectivewas to developminimally interference‐prone assays for cTnI byemploying recombinant antibody fragments. Fast 5‐ and 15‐minute assays forcTnI and D‐dimer, a degradation product of fibrin, based on intrinsicallyfluorescent nanoparticles were introduced, thus highlighting the versatility ofnanoparticles as universally applicable labels. The utilization of antibodyfragmentsindifferentversionsofthedevelopedcTnI‐assayenableddecreasesintheusedantibodyamountswithoutsacrificingassaysensitivity. Inaddition, theutilization of recombinant antibody fragments was shown to significantlydecreasethemeasuredcTnIconcentrationsinanapparentlyhealthypopulation,aswellasinsamplescontainingknownamountsofpotentiallyinterferingfactors:triglycerides, bilirubin, rheumatoid factors, or human anti‐mouse antibodies.When determining the specificity of four commercially available antibodies forcTnI,twooutofthefourcross‐reactedwithskeletaltroponinI,butcausedcross‐reactivityissuesinpatientsamplesonlywhenpairedtogether.

In conclusion, the results of this thesis emphasize the importance of carefulantibody selection when developing cTnI assays. The results with differentrecombinant antibody fragments suggest that the utilization of antibodyfragments should strongly be encouraged in the immunoassay field, especiallywithanalytessuchascTnIthatrequirehighlysensitiveassayapproaches.

Tiivistelmä

10

TIIVISTELMÄ

Sydänperäiset troponiinit (cTn, engl. cardiac troponin) I ja T ovat tällä hetkelläsuositusten mukaiset biomerkkiaineet akuutin sepelvaltimotautikohtauksendiagnosoinnissa ja riskinarvioinnissa. Viime vuosina markkinoille onkin tulluterittäinherkkiäcTn‐määrityksiä,joillaonmahdollistamitataverenkierrostacTn‐pitoisuuksia jopa yli 50 prosentilla terveistä henkilöistä. Näiden herkkien cTnI‐määritysten käyttöönotto on kuitenkin johtanut määritysten spesifisyydenlaskuun, jolloin potilaan kliininen kuva ei aina tue mitattua biomerkkiaineenmääritysrajanylittävääpitoisuutta. Spesifisyyden lasku saattaa johtuaerilaisistaanalyyttisistähäiriötekijöistä.

Väitöskirjatyön tarkoituksena oli kehittää nanopartikkelileimoihin perustuviaimmunometrisiä määrityksiä, joita voitaisiin käyttää automatisoidussavieritestaussysteemissä. Tärkeimpänä tavoitteena oli kehittää rekombinanttisiavasta‐ainefragmentteja hyväksi käyttäen cTnI‐immunomääritys, joka ei olisiherkkä erilaisille analyyttisille häiriöille. Fluoresoivien nanopartikkelien yleis‐käyttöisyys leimateknologiana todistettiin kehittämällä cTnI:lle ja fibriininhajoamistuotteelleD‐dimeerillenopeitamäärityksiä,joidenvalmistumisaikaon5tai 15 minuuttia. Vasta‐ainefragmenttien käytön avulla voitiin vähentääkäytettyjen vasta‐aineiden määriä cTnI‐määrityksessä sen herkkyyttä alenta‐matta. Lisäksi rekombinanttisten vasta‐ainefragmenttien käyttö alensimerkittä‐västimitattuja cTnI‐arvoja terveiltä ihmisiltäperäisinolevissanäytteissä,muttamyös näytteissä, joissa oli tunnettu määrä mahdollisia häiriötekijöitä:triglyseridejä, bilirubiinia, reumatekijöitä tai ihmisen anti‐hiiri vasta‐aineita.Kaupallisten cTnI vasta‐aineiden spesifisyyttä tutkittaessa havaittiin, että kaksineljästä tutkitusta vasta‐aineesta sitoutui cTnI:n lisäksi myös luusto‐lihasperäiseen troponiini I:hin, mutta ristireaktio aiheutti virheellisiä tuloksiavain silloin, kun näitä kahta vasta‐ainetta käytettiin parina samassamäärityksessä.

Väitöskirjatyön perusteella huolellinen vasta‐aineiden valinta on äärimmäisentärkeää cTnI‐määritysten kehityksessä, jotta voidaan taata määritysten cTnI‐spesifisyys.Lisäksituloksetrekombinanttivasta‐ainefragmenteillaosoittavat,ettäniidenkäyttöätulisivahvastilisätäerilaisissakaupallisissaimmunomäärityksissä‒ etenkin sellaisilla analyyteillä, joilla pyritään kehittämään hyvin herkkiämäärityksiä.

Introduction

11

1 INTRODUCTIONThehistoryofimmunoassaysgoesbackseveraldecadestothelate1950s,whenradioimmunoassay for insulin was first introduced (Yalow and Berson, 1959).Since then, many advances and changes have been introduced, the mostinfluentialbeingthedevelopmentofhybridomatechnologyinthe1970s(Köhlerand Milstein, 1975). Hybridoma technology provides a continual source ofmonoclonalantibodies(Mabs)thathavespecificitytoacertainepitope.Thishasensured their predominant position as immunoassay reagents, as well as theirutilizationasagrowing technology incancer therapies (Ribatti,2014).Not longaftertheintroductionofhybridomatechnology,intheearly1980s,theutilizationof recombinant DNA techniques resulted in the production of recombinantantibodies nearly identical to the ones obtained from hybridomas (Boss et al.,1984). During the last 20 years, automated plate‐reading systems and theintroductionofpersonalcomputersfordataanalysishavebroughtimmunoassaytechnologiestoeverydayclinicaluse.

Immunoassayinterferenceshavebeenacknowledgedfromthebeginning:alreadyin1956,Bersonetal.noticedthatlabeledinsulinwasboundunexpectedlyinthesera of patients receiving insulin treatment. Since then, it has been noted thatwheneverimmunoassaysareusedinclinicallaboratories,theyarevulnerabletodifferentanalyticalandpreanalyticalinterferencesleadingtoeitherfalsepositiveor negative results (Bolstad et al., 2013; Tate andWard, 2004). One source ofthese interferences is the antibody parts originating in the host mammal, e.g.,mouse (Kricka, 1999).Also, the currentdesire todevelop increasingly sensitiveassays for certain analytes, such as cardiac troponins (cTn), may render theassays increasingly susceptible to low‐affinity interferences, causing decreasingassay specificities and difficulties in the diagnostic process (Zaidi and Cowell,2010; Robier etal., 2014). Some of thesematrix‐related interferences could beavoidedby employing antibody fragments that canbeoptimizedby eliminatingthecomponentsmostpronetodifferentinterferences(Hudson,1998).

Oneofthetrendsintheimmunoassayfieldduringthepastfewyearshasbeentheintroductionoffastandsimplepoint‐of‐caretesting(POCT),describedaseasytouse rapid assays that do not need to be performed by trained laboratorypersonnelinacentrallaboratoryenvironment.ThemotivationbehindthisPOCTtrendhasbeenthepossibilitytotriagepatientsasearlyaspossiblealongwiththeinterest inthepromotionofhealthcareinthedevelopingcountries. In2012,theglobal in vitro diagnostics market was worth $72 billion, of which theimmunoassaymarket,excludingPOCT,covered$18.2billion.Althoughasawholetheimmunoassaymarketisshowingsignsofmaturing,POCThasbeenthefastestgrowingsectorofinvitrodiagnosticswithanannualsalesof$21.5billionin2012.(Huckle,2013.)

In this thesis, highly fluorescent nanoparticles were used to developimmunoassayspossiblyapplicabletoPOCTpurposes.Specialinterestwasfocused

Introduction

12

on cardiac troponin I (cTnI) assays, in which recombinant antibody fragmentswere used to develop rapid andminimally interference‐prone assays. Also, theversatilityofnanoparticlesasuniversallabelsforPOCTapplicationsaswellasthespecificity of commercial cTnI antibodies was studied. The following literaturereview will offer an overview of immunoassay interferences, and differentapproachesused to avoid and eliminate them.The focuswill be on, thoughnotrestrictedto,immunometricassays.

LiteratureReview

13

2 LITERATUREREVIEW

2.1 CardiactroponinIAcute coronary syndrome (ACS) that includes unstable angina and myocardialinfarction (MI) is one of the leading causes of deathworldwide. Approximately15‒20 million people present to the emergency department (ED) yearly withsymptoms indicative of ACS (Mueller, 2014). The current recommendation fordiagnosingMI includes detecting the rise and/or fall of the cardiac isoforms oftroponins I (cTnI) or T (cTnT), along with an analysis of electrocardiographicabnormalitiesandpatientsymptoms(Thygesenetal.,2012).

The troponin complex (TnI‐TnT‐TnC) functions as a regulator of musclecontraction and is found in the filaments of both cardiac and skeletal striatedmuscles (Farah and Reinach, 1995). TnI is a 21–24 kDa protein (Perry, 1999)existing as three genetic isoforms: cardiac‐specific cTnI and two differentsubforms (fast and slow‐twitch) of skeletal muscle specific troponin I (skTnI)(Hastings,1997).

The prevailing view is that cTnI is only released from cardiac muscle afterirreversibledamagetotheheart,althoughsomesuggestionsofcTnIreleaseduetoreversibleinjuryhavealsobeenmade(Bergmannetal.,2009;Hickmanetal.,2010).Aftercellmembranedisruption, thefreecytosolicpoolofcTnI(2‒6%)isreleasedinthecirculationwithinhoursafterthemyocardialdamage(Hickmanetal., 2010). This initial cTnI release is then followed by the breakage of themyofibrils causing the release of the complexed cTnI during the following 5‒7days(WuandChristenson,2013).InMI,cTnIhasmainlybeenfoundasapartofbinary (cTnI‐TnC) or ternary complexes (cTnI‐cTnT‐cTnC) and only in smallamounts as free cTnI (Katrukha et al., 1997; Giuliani et al., 1999; Bates et al.,2010).

2.1.1 ImmunoassaysforcardiactroponinIThefirstreportofcTnIasamarkerforMIwaspublishedonlyin1987(Cumminsetal.),andthefirstcommercialassayswereretailedbySanofiPasteur(Larueetal., 1993) and Dade Behring (Adams et al., 1993) in 1996. During the past 30years,cTnIandcTnThavebecomethegoldenstandardbiochemicalmarkersfordiagnosing MI, due to the first Universal Definition of MI that encouraged theassaymanufacturerstoreducethetotalimprecision(coefficientofvariation,CV)ofthecTnassaystobe10%attheMIdecisionlevelof(99thpercentile)(Alpertetal.,2000;Thygesenetal.,2012).

To enable the diagnosis of ACS as early as possible, a new generation of high‐sensitivityassaysenablingthemeasurementofcTnsinatleast50%(ideally95%)ofthereferencepopulationwithatotalimprecision(CV)of10%haveemergedthemarketduringthepastfewyears(Appleetal.,2012a).Thesehighsensitivity

LiteratureReview

14

assayshavefacilitatedsuperiorperformanceinrapidACSdiagnosis(Kelleretal.,2009;Reichlinetal.,2009)aswellasinpatientriskstratification(Giannitsisetal.,2010a;Lindahletal.,2010).

2.1.2 Point‐of‐caretestingThroughout theyears,POCThasbeendefined invariousways.Oneof themostrecentonesis:testingconductedoutsidethelaboratorysetting,nearthepatient,andbyapersonwhoseprimarytrainingisnotintheclinicallaboratorysciences(Nichols, 2013). Whatever the definition, none of them prefer a particulartechnologyoramethodoveranother.Thus,thetestingshouldbeperformednearthe patient, and facilitate a short turn‐around time (TAT), i.e., decreased timefrom the sampling to obtaining the result (Drain etal., 2014). This decrease inassayTATisconsideredtobethebiggestadvantagePOCThastooffer; thetimesavings are obtained by using whole blood instead of plasma or serum andeliminatingbloodtransportation(Drenck,2001).Resourcesbeinglimited,expertsintheWesternworldhavecometotherealizationthathealthcarebudgetscannotgrow at the rate they have been doing during the past decades. Thus,governments are actively seeking ways to lower healthcare costs (St John andPrice, 2013). POCT has been subject to animated discussion during the pastdecades,andtheevidenceforitscostefficiencyseemsstilltobelimitedandunderdebate(FermannandSuyama,2002;StJohnandPrice,2013).

It has been shown that in the Western world a substantial proportion ofhealthcare resources targeted toEDsareconsumedbypatientspresentingwithchestpains (Goodacreetal., 2005). Since themajority of thesepatients arenotdiagnosedwithACSoranotherlifethreateningcondition(Lindselletal.,2006),aneffectiveandrapiddiagnosiswouldbenefitallpartiesinvolved.POCTisexpectedto assist in meeting the ED guidelines for cardiac markers: TAT from bloodcollection to the result should be a maximum of 30 minutes, and theimplementationofPOCTshouldoccuriftheresultscannotbeobtainedinunder60minutes(Storrowetal.,2007).

A recent study by Blick (2014) reports an average TAT of 13.1min (±5.9min)with a cTnI POCT assay,when theTAT for a central laboratory cTnI assaywas48.2min (±9.0min).When savings in theTATwereprojected into annualbed‐savings,a totalof$738,000wasaccumulated.Appleetal. (2006)havereportedthatbyimplementingPOCTforthemeasurementofcTnI,thetimeapatientspentinthehospitalwasdecreasedby8%andtheoverallcostswerereducedby25%.

There are also challenges in POCT, the implementation of which requiresresourcestoacquirethedevices,aswellastotrainthestaff(Drenck,2001).TheEDstaffinthestudyconductedbyBlick(2014)hadtheadvantageofknowingthePOCT device from previous use for other analytes, thus no extra costs wereneededforthetrainingofthestaff.WhethertheminimizationofassayTAToffersaddedvalueintheactualACSdiagnosisremainsamatterofdebate.Thisisduetothe low number of studies that address the issue of whether POCT has actual

LiteratureReview

15

clinicalimpact.Ithasbeenshownthat,theoretically,theimplementationofPOCTandredesignof thecurrentpathwayscan limit the lengthofEDadmissionsandincrease the patient flow via reducing the laboratory TAT time (Storrow et al.,2008). According to the experts, one of the biggest shortcomings in the studiesinvolvingthePOCTofcTnsisthat,currently,therearenostudiesthatwouldhaveclearlydemonstrated that saving60minutes in thediagnosisofMI reduces theriskoffuturemyocardialevents(Bingisseretal.,2012).

2.2 Lanthanide‐dopedpolystyrenenanoparticlesThe European Committee for Standardization has defined nanomaterials asparticleshavingexternaldimensionsonthenanoscale, i.e., themaximumsizeofnanoparticlesshouldbeintheorderof100nm(Lövestametal.,2010).Anumberof different materials can be used to obtain nanosized particles (Grillo et al.,2014).Sincelanthanide‐dopedpolystyreneparticlesareusedasthelabelsinthisthesis, thefollowingsectionshortlydescribeslanthanidechelatetechnologyanditsusewithnanosizedpolystyreneparticles.

2.2.1 LanthanidechelatetechnologyLabelsconstructedtoincludechelatedlanthanideions(Eu3+,Tb3+,Dy3+,Sm3+)canbeexcitedwithcertainwavelengthsoflight,andafteraseriesofenergytransfersaluminescencesignalfromthelanthanideioncanbemeasured.Unliketraditionallabels, lanthanide luminescencehasrelatively longdecay times,betweenµsandms range, enabling its use in time‐resolved mode and resulting in minimalinterference frombackground fluorescence. Lanthanide labels also have a largeStokesshift,whichmeansthereisnooverlapbetweentheexcitationandemissionspectra.Therefore,self‐quenchingisnotaproblem,asiscommonlythecasewithtraditional labels. (Hagan and Zuchner, 2011) The introduction of lanthanidechelatetechnologyinthe1980s(SoiniandKojola,1983;Mukkalaetal.,1989)andits commercialization have enabled the development of wide range ofimmunoassays and DNA hybridization assays (Diamandis and Christopoulos,1990).

2.2.2 Productionoflanthanide‐dopedpolystyrenenanoparticlesAttachingseveralchelatesintoasinglebiomoleculehasbeenreportedtoassistinthedevelopmentofsensitiveassays(Diamandisetal.,1989),butitcanultimatelycause increases in theassaynonspecificbinding (Laukkanenetal., 1995). Sinceonly a number of individual chelates can be attached to a single protein, thelanthanidechelateshavebeenincorporatedintoashellprotectingthelabels.

The polymerization of styrene monomers is a well‐recognized method forachieving nanoscale particles (Piskin et al., 1994). One way of manufacturingnanoparticles doped with lanthanide chelates is their preparation in aqueoussolutions.Asthechelatesarehydrophobic,theyareforcedintoalesshydrophilicenvironmentinthepolymercapsules(Huhtinenetal.,2005).Over30,000ofthese

LiteratureReview

16

differentchelatescaneffectivelybe incorporatedintoasinglenanoscaleshell toproduce a particulate label, a so‐called nanoparticle, that does not affect theindividual fluorescencepropertiesof thechelates,andevenasingleparticlecanbe detected in a solution (Härmä et al., 2001; Soukka et al., 2001b). Thecomposition of the final particle solution can be modified by using differentamounts of co‐polymers in the polymerization reaction. These co‐polymersincludeacrylicormethacrylicacid,whichproducecarboxylgroupsontheparticlesurface(ShirahamaandSuzawa,1984).Bytheadditionoftheseco‐polymersthecolloidal stability and functionality of the particles can be altered (Lück et al.,1998). In the end, the particles are separated from the excess of the label(Matsuyaetal.,2003).Approacheswithsimultaneouspolymerizationandchelateencapsulation have also been introduced (Chen et al., 1999; Tamaki andShimomura, 2002). To prevent the chelate from leaking out of the polystyreneparticle,Hakalaetal. (2006)havedescribedamethodforcovalently linkingthechelate into the particle shell. Commercially available lanthanide‐dopedpolystyreneparticleswithmodifiedsurfacesarealsoreadilyavailable.

2.2.3 FunctionalizationofpolystyrenenanoparticlesThe simplest way of attaching the target molecule to the particles is physicaladsorption (Dezelić et al., 1971), which is a reversible process. The adsorbedmolecule can partially desorb due to pH changes (Ortega‐Vinuesa andHidalgo‐Alvarez, 1995), detergent addition, or displacement by anotherprotein (Baleetal.,1989).

The functional groups formedduring the polymerization reaction are conveyedonto the particle surface to enable an efficient covalent conjugation ofbiomolecules with the particles (Bale Oenick et al., 1990). Therefore, specificcoupling schemes are available for different functional groups (Brinkley, 1992;Holmes and Lantz, 2001). The utilization of polystyrene lanthanide‐dopednanoparticleshavebeendemonstratedinanumberofdifferentbioaffinityassaysforprotein(Soukkaetal.,2001b),nucleicacids(Huhtinenetal.,2004),andvirus(Valanneetal.,2005)diagnostics.Traditionally,thebioconjugationofwidelyusedcarboxylmodifiedparticlesisperformedbyusingN‐(3‐dimethylaminopropyl)‐N’‐ethylcarbodiimide (EDC) ‐chemistry.EDCactivates thecarboxylicgroupson theparticle surface and enables their reaction with primary amine groups on thebindermolecules (Figure1).The incorporationofN‐hydroxysuccinimide (NHS)orN‐hydroxysulfosuccinimide(sulfo‐NHS)yieldsamorestableintermediatethanEDCalone(Starosetal.,1986;Griffinetal.,1994).

LiteratureReview

17

Figure1. Schematic presentation of the activation chemistry of carboxyl‐modified particles. EDCactivationoftheCOOH‐groupyieldsanunstableintermediatethatcandirectlybeusedtocouplethedesired protein or other primary amine‐containing molecule to the particle. Some proteins areinactivatedfromdirectexposuretoEDC,andtheadditionofNHSorsulfo‐NHSformsamorestableintermediate, thus enabling washing steps and the removal of the unreactive EDC prior to thecouplingofthebindermolecules.(AdaptedfromGriffinetal.,1994).

2.2.4 ImmunometricassaysutilizingpolystyrenenanoparticlesSoukka et al. (2001a) have demonstrated that by optimizing the amount ofactivation reagents and the protein amount in the coating reaction, up to 200activebindingsitescanbecreatedontoasingle107nm(diameter)particle.Thiscausesan increase in thebindingareaof the labelandenhancestheassociationrate of the detection antibody used (Soukka et al., 2001a). This avidity‐basedenhancementofassaysignalassistsinthedevelopmentofhighlysensitiveassayswith simple test designs (Härmä et al., 2001; Soukka et al., 2001b). Thus,lanthanide‐doped polystyrene particles have enabled the development ofsensitive assays for different protein and viral analytes (Soukka et al., 2003;Valanneetal.,2005;Järvenpääetal.,2012).

2.3 ImmunoassayinterferencesThe concept of immunoassay interference is almost as old as immunoassaytechnology itself.The firstsubstantialreportsof immunoassay interferencedateback to the 1970s and 1980s (Sgouris, 1973; Hunter and Budd, 1980).

LiteratureReview

18

Nevertheless, immunoassay interferences are frequently reported with newassays,aswellaswithassaysthatarewidelyusedintheclinical field(Xuetal.,2013; Lippietal., 2014).Over the years, false immunoassay results have led todetrimental misdiagnosis and unnecessary treatments of patients (Rotmenschand Cole, 2000; Ballieux et al., 2008). The effects of immunoassay interferencedependon the interference type, aswell as themethod and analyte used. Theycauseeitherfalsepositiveorfalsenegativeresults,butthesecanbeobservedwithone method and be corrected with an assay from a different manufacturer(Janssenetal.,2014;Robieretal.,2014).

2.3.1 InterferencemechanismsofcirculatinghumanantibodiesAntibodiesused indifferent immunoassays sharea commonbasic structure.AnimmunoglobulinclassG(IgG)monoclonalantibody(Mab)consistsofalightchainandaheavychain.Inthelightchain,thereareconstant(CL)andvariabledomains(VL),whereasthelargerheavychainconsistsofavariabledomain(VH)andthreedifferent constant domains (CH1, CH2 and CH3). Domains present in the so‐calledfragment antigen binding (Fab) ‐region consist of VL, CL, VH, and CH1. Fragmentcrystallizable(Fc)‐regioncontainstheremainingconstantheavydomainsCH2andCH3.(Lipmanetal.,2005)ThebasicstructureofanIgGclassantibodycanbeseeninFigure2.

Figure2.ThebasicstructureofanIgGclassantibody,commonlyusedinimmunometricassays

Afalsepositiveimmunoassayresponseisobtainedwhenacirculatinginterferinghuman antibody bridges the capture and detection antibodies without thepresenceofananalyte.Negativeinterference,however,canbedetectedwhentheinterferingantibodyblocksthebindingsitesoftheassayantibodies(oneorall),sothatnobindingoftheanalytecanoccur(Figure3A).(Kricka,1999)

LiteratureReview

19

ThemostcommontypesofhumanantibodiesbindingdifferentpartsoftheassayantibodiesarepresentedinFigure3B.Humananti‐isotypeantibodiesbindtotheFc‐regionof,forexample,murineIgGantibodies(mIgG)(Csakoetal.,1988;Vaidyaand Beatty, 1992; Bolstad et al., 2011). Furthermore, interfering circulatingantibodies can be directed against the idiotype of the antibody (Goodman et al.,1985;Kuus‐Reicheletal.,1994),oragainsttheanti‐idiotypeantibody,hencecalledanti‐antiidiotypeantibodies(Frödinetal.,1992;Reinsberg,1995).

Figure 3. A) Examples of immunometric assays for a true positive, a false positive and a falsenegative immunoassay result. B) Examples of specificities of different possibly interferingheterophilicandsimilarantibodies.

2.3.2 Heterophilic‐typeinterferenceDuring the past 40 years, different immunoassays have been reported to sufferfrom interferences caused by circulating antibodies that have affinity to assay

LiteratureReview

20

antibodies. Heterophilic and similar antibodies can express immunoglobulinclassesG,A,M,andsomewhatinfrequentlyE(Kricka,1999).Oneveryprominentfieldaffectedbyheterophilic‐type interferencehasbeen theassays fordifferenttumormarkers, includinghumanchorionicgonadotropin(hCG)(RotmenschandCole, 2000; Gallagher et al., 2010), prostate‐specific antigen (PSA)(Fritz et al.,2009), cancer antigen 125 (CA‐125) (Bertholf et al., 2002), carcinoembryonicantigen (CEA) (Bjerner et al., 2002), and calcitonin (Papapetrou et al., 2006).Different categorizations of heterophilic‐type interference can be made. In thisthesis, as suggested by Kaplan and Levinson, the term human anti‐animalantibodies(HAAAs)andtheirsubclasses,includinghumananti‐mouseantibodies(HAMAs), have been reserved for high‐affinity human antibodies that areproduced against specific immunogens in response to a known exposure ofantibodiesofforeignoriginthroughdiagnosticortherapeuticpurposes.Contrarytothis,heterophilicantibodiesareusuallymultispecific,weakantibodiesthatareproducedagainstpoorlydefinedantigenswithoutanyhistoryofspecificmedicaltreatment with animal antibodies. (Kaplan and Levinson, 1999) However, theterms HAAA and heterophilic antibodies are often used synonymously in theliterature.

2.3.2.1 Humananti‐animal,anti‐chimericandanti‐humanantibodiesDuringthepastyears,therapeuticantibodieshaveincreasinglybeenintroduced,thus resulting in reports of specific HAAAs (Hwang and Foote, 2005) and theintroduction of engineered antibodies (Presta, 2006). The most commonlydescribed HAAAs are HAMAs, since antibodies of murine origin are mostcommonly used in commercial immunoassays (Kricka, 1999). Anti‐isotypeHAMAsseemtobemorecommonthanother typesofHAMA:Lindetal. (1991)reportedthatafteranadministrationofamedicalmurineMab,29outofthe141study participants showed a positive HAMA response: 80% being anti‐isotypicandtheremaining20%anti‐idiotypic.ThereportedHAMAconcentrationsinthecirculation vary hugely, and can be as high as on the g/L scale (Moseley etal.,1988). The duration of the response can be maintained for years after theexposure to foreign Igs(Baumetal.,1994). InadditiontoHAMA,specificHAAAresponses have been observed against other animals serving as hosts forimmunotherapy.Theexistenceofbothhumananti‐rabbit(Hiemstraetal.,1988)andanti‐horseantibodieshavebeenreported(Harkiss,1984).

When the Fc‐region of the Mab is substituted with the corresponding humansequences, the human antibody response is often reduced (Hwang and Foote,2005).Nevertheless,significanthumananti‐chimericantibody(HACA)responseshavealsobeenreported(Colnotetal.,2000;Afifetal.,2010).Duringtherecentyears,humanizedantibodieswiththeonlynon‐humanpartsremainingbeingtheanalyte‐recognizing ones have also emerged.With these humanized antibodies,the immunogenicity cannotbepredicted.AhumanizedMabA33 that is used inthetreatmentofcoloncancerhasshowedahumananti‐humanantibodyresponsein66%ofthetreatedpatients(Ritteretal.,2001).

LiteratureReview

21

TheproductionofHAAAsisacomplexprocess,sincenotallantibodiesofanimalorigin are reported to cause corresponding HAAAs, or interferences inimmunoassays.PatientsinjectedwithArcitumomab,amurineFab‐fragmentusedin the diagnostic imaging of colorectal cancer, do not form HAAA‐responsesagainst the drug (Wegener et al., 2000). Furthermore, not all HAAA‐responsescause interferences in immunoassays.Hemophiliacs treatedwith porcine factorVIII,formHAAAsagainstthemedication,butnoimmunoassayinterferencesfromtheseHAAAshavebeenreported(Morrisonetal.,1993).

2.3.2.2 HeterophilicantibodiesAlthough heterophilic antibodies are generally considered to be low‐affinitymultispecific antibodies, also high‐affinity antibodies especially againstimmunoassay antibody sub‐class IgG1 have been found (Bjerner et al., 2005;Bolstad et al., 2011). Heterophilic antibody interference seems to be causedmainlybyhumanimmunoglobulinclassM(IgM)antibodies(Bjerneretal.,2005).Due to the polyvalence of IgM class antibodies, the interference from IgM classantibodiesisoftenconsideredmoreeffectivethantheinterferencefromIgGclassantibodies. Despite the current view that heterophilic antibodies are formedwithoutaknownexposuretoanimalantibodies,theyaremorecommoninpeoplethatareroutinelyhandlinganimals(Davies,2013),andtheirproductionisoftenconsideredasanantigendrivenprocess(Checketal.,1995).Sourcesproposedforthe induction of heterophilic antibodies include exposure to animals (Berglundand Holmberg, 1989) and animal‐based products (Dahlmann and Bidlingmaier,1989), blood transfusion (Davidsohn and Lee, 1966), autoimmune diseases(FalchukandIsselbacher,1976),andmaternaltransfer(Larssonetal.,1981).

Inmanycasesheterophilic‐typeantibodyinterferencecanbetracedtobeagainstantibodies from a specific origin. Heterophilic human anti‐bovine antibodies(Andersen et al., 2004), anti‐sheep antibodies (Monchamp et al., 2007), anti‐rabbit antibodies (Butch, 2000), and anti‐goat antibodies (Cavalier etal., 2009)haveallbeenreportedtocauseimmunoassayinterferences.Althoughthesetypesof immunoassay interferences are nowadays often caught before causingdetrimental effects, reports of overlooking these interferences have beenpublishedthroughouttheyears.InareportfromBerglundandHolmberg(1989),acasewaspresented,whereheterophilicantibodiesspecifictorabbitantibodiescaused falsely elevated levels of follicle‐stimulating hormone (FSH) in a patientpresentingwithinfertilityandamenorrhea.DuetothefalselevelsofFSH,aseriesofunnecessaryproceduresincludinganovarianbiopsyweredone.

IgGsfromdifferentmammalspossesssequencehomology.Therefore,heterophilicandsimilarantibodiesareknowntocross‐reactwithIgGsfromdifferentanimalspecies.Thompsonetal. (1986)showedthat theadministrationofnon‐immunesera fromdifferent animals could effectively eliminate heterophilic interferenceagainst creatine kinase muscle‐brain fraction (CK‐MB) assay antibodies. Theadditionofmouseserumabolishedtheinterferencein100%ofthesamples,when

LiteratureReview

22

the additions of sheep, cow, guinea pig, rat and rabbit serum abolished theinterference in 78%, 78%, 69%, 70% and 25% of the samples, respectively.Sampson et al. (1994), however, reported that heterophilic interference in twomalesamplescouldequallybeblockedbyusingmouse,sheep,orgoatserum,aswellasbymIgG1,mIgG2aandratIgG.

Heterophilic‐typeinterferencecanalsobecausedbyantibodiesproducedagainstmicro‐organisms. Covinsky et al. (2000) have reported on a 56‐year‐old malesuffering from Escherichia coli septisemia, who presented with falsely elevatedassay values for cTnI, thyrotropin, hCG, α‐fetoprotein, and CA‐125. SampleincubationwithmouseMabsorE.coliremovedtheinterferenceandcorrectedtheassayresults.Thus,anextremelyrestrictedIgMantibodyresponseagainstE.coliwascausingthespuriouslyelevatedassayvalues.

2.3.2.3 RheumatoidfactorsRheumatoid factors (RF) are autoantibodies (usually IgM) directed against apatient's own IgG and IgA antibodies. About 70% of the people affected withrheumatoid arthritis are found to have circulating RFs (Wolfe etal., 1991), butRFscanalsobefoundinapproximately5%ofthehealthypopulation(Nielsenetal., 2012). Originally, RFs were defined by their ability to bind to antigenicdeterminants in the antibody Fc‐region (Carson et al., 1987), but they can betargeted to other regions as well (Selby, 1999). The presence of RFs has beenconnectedtofalselyelevatedanalytevaluesinnumerousassays,includingthyroidtesting (Marteletal.,2000), tumormarkerassays (Berthetal.,2006),andcTns(Fitzmauriceetal.,1998)mostprobablyduetocross‐reactivityoftheFc‐regionindifferentanimalspecies.

2.3.2.4 Incidenceofheterophilic‐typeinterferenceThe incidence of heterophilic‐type interference is highly dependent on theimmunoassaytypeandanalyte,andvariesgreatlybetweendifferentstudies.Inastudy by Andersen et al. (2004), it was observed that human heterophilicantibodieswithspecificitytobovineantibodieswerepresentin99%ofthedonorserum samples and caused a false positive interference rate of 81% in animmunoassay for endometrial protein PP14. A recent study by Koshida et al.(2010) indicated the existence of heterophilic antibodies with anti‐mousespecificity in>10%of thesamples tested.AnolderstudybyBoscatoandStuart(1986) observed an approximately 15% interference rate in non‐blockedimmunoassays with roughly 40% of the serum samples containing significantamountsofheterophilictypeantibodies.

Theextentof the immunoassay interference incidencewasclearlyhighlightedadecadeago,whenMarksetal.(2002)measuredtheconcentrationsof74differentanalytes in samples obtained from 10 different individuals with diseasesassociated with RFs. The study was conducted in 66 clinical laboratories in 7countries.Examplesofthedifferentanalytesaffectedbyheterophilicinterference

LiteratureReview

23

can be seen inTable1. Of the 3445 results obtained, approximately 9%werefound to be falsely positive. Of those false positives, 21% (1.8% of all results)wereconsideredaspotentiallyleadingtoincorrectclinicalinterpretation.

Table 1. An example of false positive results with immunoassays in a multicenter study of 74analytes.(AdaptedfromMarksetal.,2002).

Analyte Assaysystemstested(n)

Heterophilicfalsepositives/numberofanalysesperformeda

Etiologicallyuncertainfalsepositives/numberofanalyses

performedb

Cortisol 8 0/85 17/85(20%)

CA‐19‐9 10 1/84(1.2%) 9/84(11%)

Estradiol 7 3/59(5.1%) 34/59(58%)

Myoglobin 7 28/59(48%) 0/59

TSH 14 10/249(4.0%) 4/249(1.6%)

cTnI 9 18/156(12%) 9/156(5.8%)a Samples that were reduced to reference ranges by pre‐treatment with blocking reagent, andsamplesthatwereoriginallywithinreferencerange,butwerereduced>30bypre‐treatmentwithblockingreagent.bSamplesthatwerenotreducedtoreferencerangesbypre‐treatmentwithblockingreagent.

2.3.3 AutoantibodiesagainsttheanalyteAutoantibodiesareproducedagainstautoantigenswhenthebody’sself‐toleranceis failing. Autoantibodies are produced in many autoimmune diseases, e.g,rheumatoid arthritis, but can also be found in apparently healthy individuals(Nielsenetal., 2012).Autoantibodies against theassayanalyte can causeeitherfalse negative or false positive signals in immunoassays. In thyroid testing, theprevalenceofautoantibodieshasbeenestimatedtobebetween0%and25%,andtheir occurrence has been connected to interferences in thyroid function tests(DespresandGrant,1998).Falselydecreasedvaluesofinsulincanbemeasured,iftheinsulinautoantibodiesmaskthebindingsitesoftheassayantibodies(Kimetal., 2011). Autoantibodies have been detected against various other analytes aswell, including PSA (Zisman et al., 1995; McNeel et al., 2000) and calcitonin(Dorizzi et al., 1991), but their impact on the measurement of the analytes isunknown.

A number of studies have highlighted the effects of anti‐analyte autoantibodiesagainstcTns.Theseautoantibodiescancausefalsenegativeresultsby inhibitingthecTndetectionoftheassayantibodies(Erikssonetal.,2005b),orfalsepositiveresults by stabilizing the antigen (Plebani et al., 2002). The estimates of cTnautoantibody prevalence in an apparently healthy population vary extensivelyfrom 0% (Shmilovich etal., 2007) to as high as 20% (Lappé etal., 2011). Thishugevariationispossiblyduetothedifferences inthedetectionmethods,assaysensitivities,andthenumberofstudysubjects.

LiteratureReview

24

2.3.4 OtherproteinsVarious other proteins can also cause interferences in immunoassays.When anantigen binds to a solid‐phase antibody, or an intactMab is bound to a plasticsurface,animmunecomplexisformedthatmayactivatethecomplementsystem(CarlanderandLarsson,2001;Nilssonetal.,1993).ComplementfactorC4bbindsto the Fab‐region of the antibody (Campbell etal., 1980),whichmay cause theantibodybindingsitestobeblockedbythecomplement.Ithasalsobeenreportedthat antibody subclasses show different susceptibilities to complement‐derivedinterference so that subclass IgG2 antibodies are more prone to complement‐derivedinterferencethansubclassIgG1antibodies(Börmer,1989).

Lysozymebinds proteinswith low isoelectric points, thus possibly bridging theassayantibodies(Selby,1999).Differenthormone‐bindingproteinscanalterthemeasured analyte concentration by blocking the analyte or by removing theanalyte from the assay antibodies (Tate andWard, 2004). Albumin may causeinterference as a result of a high blood concentration, or the ability to bind orreleasevastquantitiesofligands(Selby,1999).

Paraproteins, which are antibodies that do not fight against infection, arereportedtocausefalselydecreasedimmunoassayvaluesforvancomycin(LeGattet al., 2012) and thyroid‐stimulating hormone (TSH) (Loh et al., 2012).Macrohormones,whicharehormonesconjugatedwithimmunoglobulins,canalsocause assay interferences, usually by artefactually elevating assay results. Themostcommonlyencounteredmacrohormoneismacroprolactin,whichmaycausefalse diagnoses of hyperprolactinaemia (Fahie‐Wilson and Smith, 2013), butreports of macrohormone‐induced elevations of assay results have also beendocumentedwithTSH(Lohetal.,2012)andB‐typenatriureticpeptide(Janssenetal.,2014).

2.3.5 Hemolysis,icterusandlipemiaUnlike assays employing spectral and chemical techniques, the majority ofimmunoassaysarenotaffectedbyhemolysis,icterusandlipemia(TateandWard,2004).Nevertheless, certainassaysandanalyteshavebeen foundtobeaffectedbythem(JiandMeng,2011).HemolysishasbeenassociatedwithdecreasedcTnTvalues, possiblydue to thedegradationof cTnTby the releasedproteases fromhemolyzedbloodcells(Sodietal.,2006;JiandMeng,2011).Icterus,i.e.,elevatedlevelsofbilirubin,hasbeenlinkedtostatisticallysignificantdecreasesinaspecificcTnIassaywithoutunderstandingthemechanism(verElstetal.,1999;Dasguptaet al., 2001). High level of sample triglycerides and/or cholesterol, known aslipemia,however,hascausednegativeinterferenceinacompetitivetestosteroneimmunoassay(Owenetal.,2010).

LiteratureReview

25

2.3.6 Cross‐reactivityCross‐reactivity occurs, when endogenous molecules with similar or identicalepitopestotheanalytebeingmeasuredarepresent inthecirculation(KrollandElin, 1994). Therefore, cross‐reactivity requires a stereochemically permissiveenvironment at the antigen‐antibody interface (Chitarra et al., 1993). Cross‐reactivitycancauseeitherfalsepositiveorfalsenegativeresultsdependingontheepitopes present in the cross‐reactant (Sturgeon and Viljoen, 2011). Moderndigoxigenassaysshowdetrimentalcross‐reactivitywith immunoreactive factorspresent in renal failure, liverdisease, andhypertension (Dasgupta, 2006).EarlyhumanhCGassayscross‐reactedwith luteinizinghormone(ThomasandSegers,1985),butthedevelopmentofnewmonoclonalantibodieshasledtoveryspecifichCG assays with which the cross‐reactivity issues can be avoided. It has beenrecommended that, to ensure a clinically useful assessment of cross‐reactivity,cross‐reactivity should be calculated as the apparent percentual change in themeasuredendogenousanalyteconcentrationwhen thecross‐reactant ispresentinconcentrationsobservedinhealthyanddiseasedpopulations(Davies,2013).

2.3.7 PharmaceuticalpreparationsSomedrugs andherbal remedies react as cross‐contaminants.Whenmeasuringsteroids, cortisol for example, is often problematic. Berthod and Rey (1988)reported cross‐reactivities between 340%–810% for different hydrocortisonederivatives in an assay for cortisol,whereas immunoassays for cyclosporinA, adrugusedasananti‐rejectiondrugfororgantransplantation,haveshowncross‐reactivitiesupto174%forthemetabolitesoftheparentdrug(Steimer,1999).

Herbal remedies, especially used in the Easternmedicine, have recently raisedconcernsas interfering factors(DasguptaandBernard,2006).AcertainChinesemedicationpreparedfromtoadvenom(ChanSu)containingbufadienolidesthatare components structurally similar to digoxin, have been linked to both falseelevationsanddecreasesofmeasureddigoxinlevels(Dasguptaetal.,2000).

2.3.8 SampleprocessingBloodcollectionandthesampletype(serumvs.plasma)mayaffectthemeasuredresults in some assay systems and analytes (Evans et al., 2001; Krintus et al.,2014). False elevations of cTnI values have been observed in Abbott AxSym®

assay system due to fibrin present in serum samples (McClennen et al., 2003;Kazmierczak et al., 2005). Chang et al. (2003) have reported that when serumseparator tubeswere used instead of plain tubes, increased assay values for C‐reactiveprotein(CRP)weremeasured.Theproposedreasonfortheobservationwas that the gel used in the tubes adsorbed macromolecules that formedcomplexeswithCRP,andthereleaseofCRPfromthesecomplexesenhancedtheantigen‐antibodyreactioncausingincreasedCRPconcentrationstobemeasured.

LiteratureReview

26

Inappropriatesamplehandlingandstoragecanalsoaffecttheassayresults.JaneEllis et al. (2003) have reported that the hormone adrenocorticotrophin wasstableonlyfor18hoursinplasma(at+4°C),comparedtothestabilityofover120hours for 18 other hormones. Additionally, falsely elevated values due toinappropriate sampling procedures are an inherent risk that should not bedisregarded. If a sample with very high analyte concentration is followed by asample with low analyte concentration, a carryover of the analyte may occur(ArmbrusterandAlexander,2006).

2.3.9 High‐dosehookeffectIn assay systems where wide dynamic ranges are required (e.g., hCG, PSA) asaturationofthesolid‐phaseanddetectionantibodieswiththecirculatingantigencan occur. This phenomenon called high dose hook effect, although notinterferenceassuch,canleadtoapossiblemisdiagnosisofacondition(SturgeonandMcAllister,1998).Adiagrammaticrepresentationofahigh‐dosehookcanbeseen in Figure 4. High‐dose hooks aremostly seen in one‐step immunometricassays, where the simultaneous addition of capture and detection antibodiesallows them to compete for the binding of the analyte. Thus, an excess of theanalytecanminimizethenumberofanalytesboundsimultaneouslytothecaptureanddetectionantibodies(TateandWard,2004).Thehigh‐dosehookeffectcanbereducedwithbasicassaydesign:thepossiblestepsincludeincreasingtheamountofassayantibodies,dilutingthesample,andaddinganextrawashstepbeforetheadditionofdetectionantibodies(Selby,1999;DemersandSpencer,2003;Coleetal.,2001).

LiteratureReview

27

Figure4.Anexampleofapossiblehigh‐dosehookeffect.Highanalyteconcentrationsaturatestheassayantibodies,thusresultinginafalselylowassaysignalandaninvalidestimationoftheanalytevalue.

2.4 Heterophilic‐typeinterference:preventionandelimination

2.4.1 ImmunosuppressanttherapyAs noted earlier, the formation of HAAAs is a complicated process, and not allHAAAs cause assay interferences. There are few studies, where the soleadministration of a therapeutic drug is compared with a situation where it isadministrated simultaneously with an immunosuppressant drug, but theadministrationofanimmunosuppressantdoesnotalwaysdecreasetheformationofanti‐antibodyresponse(Sandbornetal.,2001).WhenagroupofbreastcancerpatientsweretreatedonlywitharadiolabeledchimericMab,theyalldevelopedaspecificHACA‐response,butwhencyclosporinwassimultaneouslydispensed,noHACA‐responseswere observed (Richman etal., 1995). In another study,wheninfliximab, a chimeric monoclonal IgG against tumor necrosis factor alpha wasrepeatedlyadministered,antibodiesagainstthedrugwereformedin75%ofthepatients if no immunosuppression therapy was included. A simultaneousadministrationofanimmunosuppressantdrugdecreasedtheincidenceofHACA‐responseto43%.(Baertetal.,2003)

LiteratureReview

28

2.4.2 ImmunoassaydesignImmunoassay design has a major effect on the probability of possibleinterferencescausedbythesamplematrix.Currently,mostimmunometricassaysaredesigned tousesubclass IgG1Mabsoriginating in themouse(Bolstadetal.,2013). One successful approach to the removal of assay interferences is theutilizationofcaptureanddetectionantibodiesofdifferinganimalorigins(Hennigetal.,2000;CarlanderandLarsson,2001).

Mabs possessing the highly immunogenic Fc‐region are prone to differentinterferences,andsincethereisnosingleblockingreagentthatcouldremoveallinterferencesfromheterophilicandsimilarantibodies(Dasguptaetal.,1999),theremoval of the Fc‐region of the antibody is considered to be one of the mosteffectiveways of removingmany interferences (Csakoetal., 1988). Thusmanycompanieshavestartedtoemployantibodyfragmentsintheircommercialassays(Bolstadetal.,2013).

2.4.2.1 EnzymaticallycleavedfragmentsThe enzymatic cleavage of Mabs can be done with different enzymes.Traditionally, the non‐specific thiol protease papain, is used to obtain Fab‐fragments,whereas F(ab’)2‐fragments are obtained by using pepsin (Figure5).(Liddell,2013)TheproductionofF(ab’)2‐fragmentswithpepsinhasprovenmoredifficult than the production of Fab‐fragments: different antibody species andsubclassesarevariablysusceptibletopepsindigestion.EspeciallyIgGsubclass2bispronetocompletedegradation,soanoptimizationofthedegradationprocessiscrucial for each antibody (Wilson et al., 2002). Because of the differences anddifficulties in pepsin and papain digestions of different IgG subclasses, alsoenzymesbromelainandficinarewidelyusedforantibodyfragmentationduetoarapidfragmentationprocessandthestabilityoftheobtainedfragments(Marianietal.,1991).

LiteratureReview

29

Figure 5. Enzymatic papain and pepsin digestions of an intact Mab to obtain Fab‐ and F(ab’)2‐fragments.

Vaidya&Beatty(1992)reportedthatoutofabout2600samplestested,81wereobserved to contain heterophilic antibodies, and out of those samples 22 %(n=18)gaveunexpectedlyhighsignals in theirassays forCK‐MB.Theyreportedthat the combinatorial use of F(ab’)2‐fragments and mIgG eliminated allheterophilic interferences in the assay. Amore recent study by Väisänen et al.(2006)reportedthatinanassayforfreehumankallikrein2,theemploymentofF(ab’)2‐fragments instead of intact Mabs resulted in a 10–100‐fold decrease infalsepositivesignals,thuspracticallyeliminatingassayinterference.InastudybyBjerner et al. (2002) an introduction of an F(ab’)2‐fragment reduced theimmunoassay interference from 4% to 0.1%. Enzymatically produced Fab‐fragments have less frequently than F(ab’)2‐fragments been applied in theeliminationofassayinterferences,butalreadyin1979itwasobservedthatwhenenzymatically cleaved Fab‐fragments were applied into an assay for insulin, atotaleliminationofRF‐interferenceoccurred(Katoetal.,1979).

2.4.2.2 RecombinantantibodiesandantibodyfragmentsTheproteolytic fragmentationof intactMabs is a labor‐intensive andexpensiveprocess, so recombinant antibody fragments expressed in E.coli have beenavailablefornearly20years(Haydenetal.,1997).Furthermore,theintroductionofdifferentrecombinanttechniquesinthe1980sand1990shasalsoenabledtheproduction of antibody fragments of fully human origin (Jespers et al., 1994).These well‐established recombinant DNA techniques can be harnessed toincorporating specific reactive groups into the fragments, thus facilitating easypurification and site‐directed labeling of the fragments, alongwith enabling the

LiteratureReview

30

continual and homogenous production of the fragments (Lyons et al., 1990;Knappik and Plückthun, 1994; Lindner et al., 1997). The use of site‐specificlabelingshouldenabletheformationofhighlyorientedantibodysurfacesandthedevelopmentof increasinglysensitiveimmunoassays(Pelusoetal.,2003). Ithasbeen reported that recombinantly produced antibody fragments have equalaffinities to correspondingMabs (Altshuleretal., 2012).Also, theproductionofrecombinant fragments inE.coli and thebiotinylationof the fragmentaspartofthe purification process do not alter the stability of the produced fragments(Erikssonetal.,2000;Ylikotilaetal.,2006).

The first report for an immunometric assay for PSA completely based onrecombinant Fab‐fragments was published in the year 2000 (Eriksson et al.,2000). Since then, anumberof assaysbasedon recombinantFab‐fragments fordifferentanalytes,e.g.,cTnI(Ylikotilaetal.,2006)andlectins(Kellyetal.,2005)havebeenpublished.As forcommercial immunoassays, the4thgenerationcTnTassayfromRoche(Mannheim,Germany)utilizedtwomurineFab‐fragmentsinasandwichformat,andthenew5thgenerationhighsensitivitymodificationoftheassay utilizes a chimeric version of the detection antibody (Giannitsis et al.,2010b).

Althoughthefirstassayutilizingsingle‐chainvariablefragments(scFv)consistingof the variable domains (VH, VL) connected by a flexible linker, was describedalreadyin1997(Kerschbaumeretal.,1997),smallscFvshavetraditionallybeenconsideredtobe lessstablethanbiggerantibodyfragments(RaagandWhitlow,1995).SomescFvsmightremainstableforlessthansixmonthsin+4°C(Huetal.,2007),causingtheiruseincommercialassaystobesomewhatlimited.Ithasbeenshown that the constant parts present in Fab‐fragments significantly increaseantibodystability(Shimbaetal.,1995).Antibodyfragmentsarelikelytoincreasetheirpopularityespeciallyinimmunoassaysdetectinguntraditionalanalytes:notbecauseofadesiretoavoid immunoassay interferences,butbecausetraditionalanimal‐based immunizationmethods cannot be applied to certain analytes, e.g.,toxic components and smallhaptens (Zhangetal., 2014;Alvarengaetal., 2014;Oyamaetal.,2013).Largeantibodylibrariescaneasilybeusedtoobtainbindersforanykindsofanalytes,makinganimal immunizationredundant(Nissimetal.,1994; Brockmann et al., 2011). Despite their possible challenges in long‐termstabilities, scFv‐fragments have successfully been employed in a variety ofdifferentimmunoassays(Warrenetal.,2005;Wangetal.,2008;Brockmannetal.,2010).

Small scFv‐fragments contain less material from the original host species thanlarger recombinant antibody fragments or enzymatically fragmented antibodiesandare, therefore, lesspronetoheterophilic interferences.Warrenetal. (2005)reportedthatoutof383samplesthatproducedafalselyelevatedresult intheirF(ab′)2‐assay,374werecorrectedwhentheusedF(ab′)2‐fragmentwasreplacedwiththecorrespondingscFv‐fragment.

LiteratureReview

31

Chimericantibodiesandantibodyfragmentsareantibodiesinwhichsequencesoftwo different origins have been combined. Usually chimeric antibodies areconsideredtobeofnon‐humanorigin,e.g.,murineantibodiesthataremodifiedtocontain human sequences. For example, the constant domains of a mouseantibody can be replaced with the corresponding human sequences, thuspotentiallyminimizingtheimmunoreactivepropertiesofthemouseparts(Liuetal., 1987). Chimeric antibodies can easily be produced and secreted inE.coli asfully active constructs (Better et al., 1988). When intact Mab and chimerichuman/mouse versions of the same tracer antibody was tested in animmunoassayforCEA,itwasnotedthatwhennoassayblockerswereused,146(29.0%)outof the503serumsamples testedproduceda falselyelevatedresultwith theMab‐version,whereasonly14 (2.8%)did sowith the chimeric system(Kurokietal.,1995).

Humanizedantibodiesaredistinctfromchimericantibodies.Theyareantibodiesof non‐human origin,whose protein sequences have beenmodified to increasetheir similarity to antibody variants produced naturally in humans. The mostcommon way of obtaining humanized antibodies is called complementaritydetermining region (CDR) grafting. When performing the humanization of amouse antibody by CDR grafting, themouse CDRs responsible for the antibodybinding are inserted into a human framework with different moleculartechniques.Thebindingaffinityofthehumanizedantibodiesmaysuffer,however,since the human framework residues may be involved in the actual bindingreaction. (Safdari et al., 2013) Nowadays, humanized antibodies are used fortreatingdifferenthumandiseases,butnostudiesoftheiruseinreducingmatrix‐related interferences in immunoassays have yet been published. A schematicpresentation of murine, human, chimeric, and humanized Fab‐fragments ispresentedinFigure6.

Figure6.Anexampleofamurine,human,chimeric,andhumanizedversionsofanFab‐fragment.

2.4.3 PhysicalandchemicaltechniquesImmunoassay interference can be removed by applying a number of differentchemicalandphysicaltechniques.Thesetechniquesareusuallytimeconsuming,

LiteratureReview

32

tedious, and often constrained to a specific group of interferences. Differentapproacheshavebeentestedonsampleswhereeitherfalselynegativeorfalselypositive results have been obtained. It must be noted though, that all of thephysical and chemical approaches are highly specific to each analyte andinterferencetype,andcannotbeuniversallyapplied.

2.4.3.1 PrecipitationTheprecipitationofinterferingproteinscanbeperformedbyusingpolyethyleneglycols(PEGs)orammoniumsulfate,whichprecipitateproteinsbyloweringtheirsolubilityinaqueoussolutions,e.g.,plasmaandserum(Bolstadetal.,2013).PEGprecipitation has been used to remove proteins, e.g., prolactinmacrocomplexesand thyroidautoantibodies frombloodsamples (Fahie‐WilsonandSmith,2013;Despres and Grant, 1998). The precipitationwith PEGs and ammonium sulfatehasbeenshowntoclearnearallIgsfromthesample,butitcanalsoabolishandirreversibly denature clinically important proteins (Bolstad et al., 2013). Allsampleproteinscanalsobeprecipitatedwithethanol(Levineetal.,1990)aswellas with trichloroacetic acid, which has been used in eliminating heterophilicinterferenceindigoxinassays(Liendoetal.,1996).

2.4.3.2 AffinityextractionDifferent affinity‐based extraction methods include the application of ionexchange,andproteinG,AandL–basedapproaches(Turpeinenetal.,1990;deJageretal.,2005).Oftheaffinity‐basedmethods,approachesusingproteinsGandAarethemostuniversallyapplicable.ProteinGbindstotheFc‐regionofallfourIgG subclasses, when protein A is known to bind to only three of the foursubclasses. A case study by Lippi etal. (2013a) demonstrated that by applyingaffinity extraction with protein A, a false positive troponin level caused byheterophilicantibodiescouldbedecreasedfrom7.9μg/Lto<0.2μg/L.Inanotherstudy, the immunodepletion of sample IgGs with protein G affinitychromatography decreased themeasured falsely elevated TSH value from 25.2mIUto2.7mIU(Rossetal.,2008).Protein‐LhashigheraffinityagainstIgMthanagainst either IgA or IgG, especially through kappa light chain interactions(ÅkerströmandBjörck,1989).InastudybydeJageretal.(2005),totalIgM,IgG,andIgAlevelswerecutby60%,whileRF‐associatedIgMlevelswerecutby89%afteranincubationof2hwithproteinL.

2.4.3.3 OtherapproachesNumerousotherapproacheshavealsobeentestedtoremovesourcesofpossibleinterference. One of these is size‐exclusion (Liendo et al., 1996), which usuallyrequires relatively high sample volumes. The use of size‐exclusion is alsorestricted because the analyte is usually highly diluted during the process(Bolstadetal.,2013).Heatingasampleat90°C(Primusetal.,1988),andsampletreatment with dithioerythritol (Müller et al., 1985) can only be used withanalytesthatcanendurehightemperaturesandantibody‐denaturingconditions.

LiteratureReview

33

2.4.4 BlockingagentsThe most commonly utilized way of decreasing assay interferences is to usespecific blocking agents as part of the assay buffers. There are several specificrequirementstheemployedblockersshouldmeet.Theblockersubstanceshould:1)inhibitthenonspecificbindingoftheassaycomponentstotheassaysurface;2)inhibit non‐specific protein‐protein interactions; 3) obtain no cross‐reactivitywiththeassaycomponents(antibodies,analyte);4)notinhibitspecificreactionsintheassayand5)toexhibitverylittlelot‐to‐lotvariation.

Basicproteinblockersincludingbovineserumalbumin(BSA),casein,andgelatincanbeused todecreaseundesirableprotein‐protein interactions (Wakayamaetal., 2008;Kenna etal., 1985). An addition of excess nonspecific Igs, e.g., bovinegamma globulin (BGG) has been proven to decrease the interactions betweenassayantibodiesand interfering factors(HunterandBudd,1980;Frengenetal.,1994).Tominimizeinterferencesfromheterophilicandsimilarantibodieswithaspecificity toassayantibodies,polyclonal IgGs fromtheantibodysourcespeciesshouldbeapplieddirectlyorasapartofserumpreparation(Princeetal.,1973;Boscato and Stuart, 1986; Boerman et al., 1990). As the most commonly usedassay antibodies are ofmurine origin, the addition of free polyclonalmIgGs innativeordenaturedformhasbeenusedtodecreaseinterferencesassociatedwithheterophilic‐typeinterferencesspecifictomouseantibodies(BoscatoandStuart,1988; Primus et al., 1988). The effects of these polyclonal mIgG preparationsoriginating in a number of different animals are based on the fact that theycontain all possibly cross‐reactive allotypic and idiotypic epitopes of mIgG(VaidyaandBeatty,1992;Reinsberg,1996).

Denatured antibodies have proven to be more potent as blockers than nativeantibodies(Bjerneretal.,2002).Oneplausiblereasonforthisisthataggregatedantibodies are more prone to form stable and intricate complexes with theinterfering antibodies. When aggregated, several potent parts are in closeproximity,thusprovidingamorealluringtargetfortheinterferingantibodiestobindto(Bolstadetal.,2013).

Several different specific blocking reagents with distinct specificities andstructures are commercially available (Table 2). The advantage of specificantibodies is that they are high‐affinity antibodies (109 L/mol vs. 106 L/molunspecific blockers), allowing them to be used in low concentrations (Kricka,1999).

LiteratureReview

34

Table2.Examplesofcommerciallyavailablespecificblockingreagents.

Tradename Consistency Producer Reference

HeterophilicBlocking

Reagent(HBR)

Monoclonalmouseanti‐humanIgM

Scantibodies,(SanteeCA)

(Kurokietal.,1995;Sosoliketal.,1997)

MAB33(MAK33)

MonoclonalmIgG1Roche

MolecularBiochemicals

(Mössneretal.,1990;Bjerneretal.,2002)

PolyMAB33(&otherMAB33familyofproducts)

PolymericmonoclonalmIgG1/Fab

RocheMolecular

Biochemicals(Kricka,1999)

ImmunoglobulinInhibiting

Reagent(IIR)

IgGandIgM(frommultiplespecies)withhighaffinity

toHAAA

BioreclamationInc,(East

Meadow,NY)

(Nicholsonetal.,1996;Reinsberg,1998)

SuperChemiblock

IgswithhighaffinitytoRF/HAMA

Millipore,(Danvers,MA)

(Waterboeretal.,2006;vanGageldonketal.,2011)

HeteroblockMixtureofactiveand

passiveblockingreagents

OmegaBiologicalsInc.(Bozeman,MT)

(Hueberetal.,2007)

HeterophilicBlockingTube

(HBT)

Lyophilizedbindersagainstheterophilicantibodies(500µlsample/tube)

Scantibodies (ButlerandCole,2001)

IntheoriginalstudyonMAK33,Mössneretal.(1990)reportedthatwhen10,000serasampleswereanalyzedina1‐stepassayforCEA,72sampleswerenoticedtocontainheterophilicinterferences.Outofthe72samples,7couldnotbecorrectedwith the addition of non‐immune mIgG, but showed normal values with theadditionofMAK33.InastudybyBjerneretal.(2002)immunoassayinterferencewas observed to be 4%, but when heat‐treated MAK33 was applied, theinterferenceincidencedroppedto0.86%.

Both IIR and HBR have shown to be specific blocking agents in HAMAinterferences observed in patients that have had cancer treatments withtherapeuticantibodies(Nicholsonetal.,1996;Sosoliketal.,1997).InastudybyReinsberg,IIRblockedHAMAinterferencesinmorecasesthanMAK33,butnotinasmanycasesaspolyclonalnon‐immunemIgG.Theproposedreasonforthiswasthat HAMA is usually directed against the allotypic and cross‐reactive idiotypicantibody determinants, thus preventing the blocking effect of reagents such asMAK33designedtoblockunspecificheterophilicinterferences.(Reinsberg,1998)However, both IIR andMAK33 have also proven their utility in blocking otherthan specific HAAA interferences. In an assay for CEA, the addition of HBRdecreasedheterophilic‐typeinterferencefrom29%to4.8%,whentheadditionofnormalmouseserato1%offinalassayvolumedecreasedtheinterferenceonlyto

LiteratureReview

35

6.2% (Kuroki et al., 1995). In another study, the addition of IIR decreased theinterference in serumsamples from28% to0%and12% to0% in rheumatoidarthritis and osteoarthritis patients, respectively (DeForge et al., 2010). AnalternativetoHBRistheuseofHeterophilicBlockingTubes(HBTs). Ithasbeenreported that in 11 samples that were observed to give falsely elevated hCGvaluesandweretreatedwithHBTs,9werefullyblockedandeventheremaining2wereobservedtohave>50%reducedanalytevalues(ButlerandCole,2001).

Ithasbeendemonstratedthat theadditionof3μg/mLofHeteroBlock inserumdiluent yielded effects comparable to immunoglobulin depletion by protein L‐sepharose precipitation (Hueber et al., 2007). The use of Super Chemiblok hasmostly been restricted to multiplex bead‐based fluorescent immunoassays(Luminex technology), where it has been noted that the incubation of sampleserum with Super Chemiblok significantly decreases the unspecific humanantibodybindingtotheassaybeads(Waterboeretal.,2006).

Alltheavailableblockersserveanimportantpurpose,buttheirapplicabilityishighlycase‐specific.Hence,theadditionofdifferenttypesofblockerscanbeusedtoidentifythetypeofinterferenceinaspecificsample.Often,acombinationofdifferentblockingapproachesisneededtoobtainthedesiredresult:deJageretal.(2005)observedareducednon‐specificbindingandmoreaccurate recovery ratesof cytokines,whenusingtheproteinL‐sepharoseinacombinationwith10%rodentserumasablocker.Thiscombinationtreatmentresultedin89%depletionofplasmaIgMRFs.

Whena falsepositiveresult isobtained thatdoesnotcorrelatewith theclinicalpicture of the patient, the recommended procedures include linearity studies,recovery experiments, and sample treatment with HBTs, or an addition ofblocking reagents with re‐measuring the sample with an alternate assay(SturgeonandViljoen,2011;Ismail,2007).Astherateofassayinterferencevariesgreatly between different assay types and analytes, the comparison of a set ofdifferent blockers is often impossible (Marks, 2002). The fact that commercialassayscontaincertainblockers,cutsbacktheinterferenceproblembutdoesnottotally abolish it: when 21,000 sampleswere screenedwith routine testing forthyroid function, 7 patients (0.03%) showed falsely increased analyte values(Wardetal.,1997).Ithasalsobeenreportedthatnotallblockingagentsreduceinterferencesasexpected,andsometimestheycanevenincreasefalselyelevatedanalyte values with certain assays (Koshida et al., 2010; van Gageldonk et al.,2011).Differentscreeningtechniqueshavealsobeentested,buttheiruse isnotcurrentlywarranted(Emersonetal.,2003).Ithasbeenproposedthatinsampleswith a high likelihoodof false positives (e.g., tumormarkers) itwouldbemorecost‐efficient to add a blocker to all tests rather than re‐measure all positiveresults(Bjerneretal.,2012).Asdifferenttypesofimmunoassayinterferencesareregularlyencounteredintheclinicallaboratories,differentandnovelapproacheseliminating them should be thoroughly studied. It is also important toacknowledge the possible presence of these interferences and encouragecommunicationbetweenclinicallaboratoriesandphysicians.

Aimsofthestudy

36

3 AIMSOFTHESTUDY

The overall aim of this thesis was to develop sensitive, fast, and minimallyinterference‐proneimmunoassaysforcTnIthatcouldbeutilizedinPOCTsettings.ForthispurposeintrinsicallyfluorescentEu‐dopednanoparticleswereemployedas labels, recombinantantibody fragmentswereutilizedwhenpossible, and thespecificityofdifferentcTnI‐specificantibodieswasstudied.

Morespecifically,theaimswere:

I Toillustratethatintrinsicallyfluorescentnanoparticlescanbeuniversallyemployedin immunoassaysdesignedforPOCTapplications,whethertheprimaryassayrequirementissensitivityorawidedynamicrangearoundapredeterminedanalyteconcentration.

II To study the effects that different antibody molecular forms have onmatrix‐relatedinterferencesinacTnIresearchimmunoassay.

III TopresentanovelheterogeneousresearchimmunoassayforcTnIandtoemphasize the effects the removal of antibody Fc‐region andchimerization of Fab‐fragment have on matrix‐related interferences insamplescontainingknownamountsofpossiblyinterferingsubstances.

IV ToevaluatetheskTnIcross‐reactivityoffourdifferentcTnIantibodiesinfour different versions of a cTnI research assay possessing identicalepitopespecificitydeterminedbypeptidemapping.

SummaryofMaterialsandMethods

37

4 SUMMARYOFMATERIALSANDMETHODS

A summary of materials and methods is described here. More detailedinformationcanbefoundintheoriginalpublications(I–IV).

4.1 SamplesThe summary of sample panels used in the study is described inTable 3. Allsampleswerecollectedaccordingtonormallaboratoryroutinesandwerestoredat ‐20°C for short‐term preservation or at ‐70°C for long‐term preservation.Informed consent was obtained from all participants and the study protocolswere approved by the local ethics committees. All the study protocols were inaccordancewiththeHelsinkideclarationasrevisedin2006.

SamplesfromOuluUniversityhospital(I,III,IV)werecollected,andthereferenceanalytevaluewasmeasured,asapartofroutineanalysis.Thereafter,thesampleswere frozen and shipped to the University of Turku without any identificationlabelsorfinaldiagnoses.Routinein‐houseplasmasamplesthatwerecollectedattheUniversity of Turku, Department of Biotechnology (UTU/BT)were used fortheoptimizationprocessoftheimmunoassays,aswellasforassessingthematrix‐related interferences of the different assay versions (II‒IV). All the normalsamples used in cTnI assay versions were tested for cardiac specificautoantibodies(Erikssonetal.,2005a;Savukoskietal.,2014).

Table3.Descriptionofthesamplesusedinthestudies.Samplesobtainedfromoneindividualmayhavebeenusedinseveralstudies.

Collectionplace Studypopulation Matrix Numberofsamples

Publicationanduse

OuluUniversityHospital

Leftover Citrateplasma 65 I D‐Dimermethodcomparison

UTU/BT Healthyvolunteers

Heparinplasma/serum

32(I)/64(III)/3(IV)

II,III,IVAssayinterference

studies/recovery&reference

populationstudiesOuluUniversity

HospitalLeftover Heparinplasma 265(III)/

101(IV)III,IVMethodcomparison

ProMedDx(Norton,MA,UnitedStates)

Triglyceridecontaining

Heparinplasma 12 IIIAssayinterferencestudies

Bilirubincontaining

Heparinplasma 12

RFcontaining Serum 12

HAMAcontaining

Serum 3


38

4.2 Reagents

4.2.1 AntibodiesThe antibodies used in the studies are listed in Table 4. According to themanufacturer, Mab‐FDP14 recognizes the degraded forms of human fibrin andfibrinogen,butnotintactfibrinandfibrinogen,andF(ab’)2‐8D3recognizestheD‐dimer domain contained in the cross‐linked degradation products of humanfibrin.AlltheusedcommercialcTnIantibodieswereclaimedtobecTnIspecificbytheiroriginalmanufacturers.

Table4.Antibodiesusedinthestudies.

Clonename

Specificity Molecular form Epitope(aminoacidresidues)

Manufacturer Publication

19C7 cTnI Mab 41‒49 HyTest Ltd.(Turku,Finland)

II‒IV

8I7 cTnI Mab 169‐178a(137‒148)

International PointofCare(IPOC,

Toronto,Canada)

I‒IV

9707 cTnI Mab/F(ab')2/Fab/cFabb

190‒196 Medix Biochemica(Kauniainen,Finland)

I‒IV

625 cTnI Mab 169‒178 HyTest Ltd. IV

MF4 cTnI Fab 190‒196 HyTest Ltd. IV

11N11 cTnI cFab 160‒179 UTU/BT III

FDP14 D‐dimer Mab ‐ Biokit (Barcelona,Spain)

I

8D3 D‐dimer F(ab')2 ‐ Biokit I

a(Vylegzhaninaetal.,2013)bcFab,chimericfragmentantigenbinding

TheF(ab’)2‐9707usedintheassayswasproducedfromMab‐9707byenzymaticdigestionwithbromelain(ID‐Diluent1;Diamed,Cressier,Switzerland)(Väisänenetal.,2006).RecombinantFab‐9707andFab‐MF4werecloned fromhybridomacell lines of the corresponding Mabs at the Department of Biotechnology,University of Turku. The cFab‐11N11 used was developed at the University ofTurku Biotechnology department. Both cFab‐11N11 and cFab‐9707 weremodifiedasmouse/humanchimericantibodyfragmentsbyreplacingtheconstantparts of the light and heavy chains originating from mouse with thecorresponding human antibody sequences. Unlike cFab‐11N11, Fab/cFab‐9707and Fab‐MF4 were designed to contain an unpaired cysteine residue in the C‐terminal end of the heavy chain peptide to facilitate their site‐specificbiotinylation.


39

4.2.2 Calibrators

4.2.2.1 D‐dimerD‐dimer is a degradation product of fibrin that is liberated into circulation incertainpathological conditions,e.g., venous thromboembolism(VTE),whichcanbeclassifiedasdeepveinthrombosisandpulmonaryembolism(Bounameauxetal., 2010). D‐dimer is a term used for several cross‐linked fibrin degradationproducts(FDPs)ofvarioussizes,allcontainingtheD‐dimerepitope.ThesizesoftheFDPsvaryfrom200kDaoffreeD‐dimerfragmenttoover2300kDa(Walkerand Nesheim, 1999). Although present in VTE, D‐dimer is not VTE specific;slightlyelevatedD‐dimerlevelscanalsobemeasuredinseveraldifferentillnessesandmedical conditions, including infections, traumas, collagenandautoimmunediseases, atherosclerosis, and even in uncomplicated pregnancies in healthywomen(Eichinger,2005;Adametal.,2009).ThisincreaseinthebaselinelevelofD‐dimermaybeduetoaconditioninwhichaformationof fibrinoccurs,butnoevident thrombus is formed (Brenner et al., 1990). Nonetheless, the risk forthrombosis is increased inmanycases (Heitetal.,2001; Ismaetal.,2009).Thewidelyusedclinical cut‐off levelofD‐dimer is500ng/mL fibrinogenequivalentunits(FEU)(Bounameauxetal.,2010).

TheD‐dimercalibrationmaterialwaspurchasedfromBiokit,andwasobtainedaspartially purified D‐dimer from human fibrin that was digested with humanplasmin.Thecalibratorwasdilutedinabuffercontaining137mmol/LNaCl,2.7mmol/LKCl,10mmol/LNa2HPO4,2mmol/LKH2PO4,and10g/LBSA(pH7.2).

4.2.2.2 TroponinIHuman cTn (native, tissue derived, cTnI‐cTnT‐TnC complex) and native skTnIwerepurchasedfromHyTestLtd.TheskTnIwasoriginallydilutedwithurea/Trisbuffer, pH 7.5 (7 mol/L urea, 5 mmol/L ethylenediaminetetraacetic acid, 15mmol/L mercapthoethanol, 20 mmol/L Tris). All TnI calibrators used in theassayswerepreparedbydilutingthemwithTris‐basedbuffer(50mmol/LTris,150mmol/LNaCl,0.5g/LNaN3)supplementedwith75g/LBSA(pH7.75).

4.2.3 OtherreagentsandassaybuffersAllimmunoassayswereconductedinlow‐fluorescentstreptavidin‐coated96‐wellmicrotitration wells (Kaivogen Oy, Turku, Finland). In this study, twoimmunoassay approaches, methodologically slightly different from each other,werestudied.Thepreincubationofthesampleandparticles(I)wasconductedonmicrotitrationwells(Nunc,Roskilde,Denmark)coatedwith5g/LBSA.Allsolid‐phaseantibodydilutionswerepreparedtoassaybuffer(KaivogenBufferSolutionRed, Kaivogen Oy) and microtitration wells were washed with Kaivogen washsolution (KaivogenOy). In study I, particlesweredilutedwithparticlebufferA,whichconsistedofassaybuffersupplementedwith0.1g/LnativemIgGand0.05g/L denatured mIgG. In all the versions of the cTnI research assay (no


40

preincubation step, II–IV) the particles were diluted with particle buffer Bcontaining:37.5mmol/LTris,pH7.75,500mmol/LNaCl,0.4g/LNaN3,0.6g/LBGG, 25 g/L BSA, 50 g/L D‐trehalose, 0.8 g/L nativemIgG, 0.05 g/L denaturedmIgG, 2 g/L casein, and 37.5 IU/mL heparin, modified from (von Lode et al.,2003). In the assay based on chimeric antibody fragments (III), 0.15 mol/Lbiotinylated5kDaPEGwassupplementedintotheparticlebufferB.

4.3 Preparationofassayreagents

4.3.1 Carboxyl‐modifiednanoparticlesEuropium‐fluorescentFluoro‐MaxTMpolystereneparticles(SeradynInc,IN)wereemployedthroughoutthisstudy.TheparticlesizeemployedforpublicationsI,II,andIVwas107nm,whereasthesizeusedforpublicationIIIwas100nm.NHS‐and EDC‐chemistries at +23 °C and under vigorous shaking were used in allparticlecoatings.

The107nmparticles(e.g.,1.5x1012units)werewashedwithaphosphatebuffer(10mmol/L, pH 7.0) usingmicrofiltration centrifugal devices (300 kDa cut‐off,Pall,MA),andtheywereactivatedfor15minuteswith10mmol/Lsulfo‐NHSand0.75mmol/L EDC. The covalent coupling of detection antibodies (Mab‐9707, I;Mab‐817,I,II,andIV;Mab‐FDP14,I;F(ab’)2‐8D3,I;MAb‐625,IV)wasperformedin 10mmol/L phosphate buffer supplementedwith 100mmol/L NaCl (pH 8.0,2h).Theamountsofantibodyusedvariedbetween0.5and1.0g/L.Theblockingoftheremainingactivesitesontheparticles(10g/LBSA,15minutes),aswellasthefinalstorageoftheconjugatedparticleswasdoneinaTris‐basedbuffer(e.g.,10 mmol/L Tris, 0.5 g/L NaN3, pH 8.5). For particle storage, the buffer wassupplementedwith2g/LBSA.

The100nmparticle coating (III)wasexecuted in a2‐(N‐morpholino)ethane‐sulfonicacid(MES)buffer(20mmol/L,pH6.0).First,theparticles(1.35x1012units)werewashedbyusingmicrofiltrationcentrifugaldevices (300kDacut‐off),afterwhichtheywereactivatedfor15minuteswith8mmol/Lsulfo‐NHSand 1.5 mmol/L EDC. Then, cFab‐11N11 (0.5 g/L) was let to adsorb to theactivated particles for 30 minutes (20 mM MES, 100 mmol/l NaCl, pH 6.0.).Afterthe30‐minuteadsorption,thecovalentcouplingoftheactivatedparticlesand cFab‐11N11was generatedby increasing the reactionpH to 8.0with1MNa2CO3/NaHCO3 ‐buffer (pH 9.3), and the reaction was continued for anadditional30‐minuteperiod.Theblockingoftheremainingactivesitesontheparticles(10g/LBSA,30minutes),aswellasthefinalstorageoftheconjugatedparticleswasdone inTris‐basedbuffer(25mMTris,150mmol/LNaCl,1g/LNaN3, pH 7.8). For particle storage, the buffer was supplementedwith 2 g/LBSA.

Beforethefirstinstanceofuse,theparticlesolutionsweremixedandsonicated.After that, the solutions were centrifuged to remove noncolloidal aggregates


41

(350 x g, 5min). Theparticle concentrationsweredeterminedbydiluting theparticles with 0.1% (w/v) Triton X100 solution, and by comparing the time‐resolvedfluorescenceofthepreparationtoaparticlecalibratorpreparedfromuncoatedparticlestock.ThemeasurementswereconductedwithVictorTM1420multilabelcounter(PerkinElmerLifeandAnalyticalSciences,WallacOy,Turku,Finland)

4.3.2 Labelingofsolid‐phaseantibodiesMab‐FDP14 and F(ab’)2‐8D3 (I) were biotinylated to free NH2‐groups with EZ‐Link®NHS‐Chromogenic‐Biotin (ThermoScientific,Waltham,MA,UnitedStates)for60minutesat+23°Cwhileprotectedfromlight.Mab‐FDP14(1.0mg/mL)wasbiotinylated in 50 mmol/L Na2CO3/NaHCO3 –buffer supplemented with 150mmol/lNaCl(pH9.6)witha10‐foldmolarexcessofthelabelingreagent,whereasF(ab’)2‐8D3(1.0mg/mL)wasbiotinylated in100mmol/Lphosphatebuffer(pH7.2) supplemented with 150 mmol/L NaCl with a 5‐fold molar excess of thelabelingreagent.

Mab‐19C7(1.5mg/mL,II–IV)andMab‐9707(0.6mg/mL,I–II)werebiotinylatedto free NH2‐groups with a 10‐fold, Mab‐8I7 (1.4 mg/mL, I) with 20‐fold, andF(ab’)2‐9707(0.9mg/mL,II)witha30‐foldmolarexcessofbiotinisothiocyanate(UniversityofTurku,Turku,Finland)(Erikssonetal.,2003).Recombinantmouseantibody fragments Fab‐9707 (II) and Fab‐MF4 (IV) were biotinylated site‐specifically to C‐terminal free cysteine as a part of their production andpurification process with EZ‐Link® Maleimide‐PEG2‐Biotin (Thermo Scientific)(Ylikotilaetal.,2006).

Afterthebiotinylationreactions,alltheantibodieswerepurifiedtwicewithNAPgel filtration colums (GE Healthcare Life Sciences, NY, United States) into 50mmol/LTris‐HCl(pH7.75),containing150mmol/LNaCland0.5g/LNaN3.Thelabeledantibodieswerestabilizedwith1g/LBSAandstoredat+4°C.

4.4 Immunoassays

4.4.1 Immunoassaywithpreincubationstep(I)Alltheutilizedimmunoassayswereperformedasheterogeneousnon‐competitiveimmunoassay format. In publication I, an additional preincubation step wasemployed in order to simulate simple microfluidic chip processing, where thesampleanddetectionantibodiesareintroducedpriortotransferringthemtothesolid‐phasesurface.TheassaysutilizedantibodypairsMab‐FDP14‒F(ab’)2‐8D3(D‐dimer) andMAb‐9707‒Mab‐8I7 (cTnI) so that both antibodies in thepairsweretestedassolid‐phaseanddetectionantibodies.

The preincubation of the particles (coated with Mab‐FDP14, F(ab’)2‐8D3, Mab‐9707,orMab‐8I7)andsamplewasdonesothat12µlofsampleorstandardwith1.44x109(D‐dimer)or1.8x108(cTnI)antibody‐coatedparticleswereaddedin


42

48 µl of particle buffer A alongwith 0–25 µg/mL free solid‐phase or detectionantibodies to BSA‐coated wells. After a preincubation of 30 seconds (+23 °C,shakingat900rpm),50µlofthesolutionwastransferredtoafreshsolid‐phasesurface.

The solid‐phase surface was constructed on streptavidin‐coated wells byincubating120ngofbiotinylatedsolid‐phaseantibodyMab‐FDP14,F(ab’)2‐8D3,Mab‐9707,orMab‐8I7in60µLofassaybufferat+23°Cwithshakingat900rpmfor 60 minutes. After the bio‐incubation, the wells were washed with washsolution and the preincubated sample‐particle solution was added into thewells.

The assaywellswere incubated for either5‐ or15‐minutes (+23 °C, shaking at900rpm)beforewashingwithwashsolutionandadirectsurfacemeasurementofthelonglife‐timefluorescenceofthenanoparticleswithVictor™1420MultilabelCounter. The 15‐minute assay incubation was employed only in methodcomparison. The basic principle of the developed two‐site immunoassay withpreincubationcanbeseeninFigure7.

Figure 7. The basic principle of an immunoassay with a preincubation step used forimmunoassaydynamicrangeextension(I).AD‐dimerassayusingMab‐FDP14asthesolid‐phaseand F(ab’)2‐8D3 as the detection antibody attached to nanoparticles is used as an example.Adaptedfrom(I).


43

4.4.2 ImmunoassaysforcardiactroponinI(II–IV)All the different cTnI assay versions used in the original publications II–IVemployed identical antibody configuration and assay epitopes as evidenced bypeptide mapping. The assay epitopes were designed to employ an antibodyconfiguration,whereinhibitingeffectsofcirculatingautoantibodiesmainlyboundto the central part of cTnI (amino acid residues 30–110) would be minimal(Eriksson et al., 2005b). The immunoassay configuration utilized in all assayversionsisshowninFigure8.

Figure8.Aschematicpresentationof theantibodyconfigurationandepitopespecificitiesused inthedifferentversionsofaresearchassayforcTnI(II–IV).ThebarrepresentsthelinearaminoacidsequenceofcTnI.TheN‐terminalregionofcTnI(aminoacidresidues1–29)isshownindarkgrey,thecentralregioninwhiteandtheC‐terminalregion(aminoacidresidues111–209)inlightgrey.Circlesindicatetheapproximateepitopesofthedifferentantibodiesused.

All theassayswereperformedso that fixedamountsofbiotinylatedsolid‐phaseantibodieswere incubated in50µLof assaybuffer at +23 °C for60minutes instreptavidin‐coatedwells.Aftertheincubation,thewellswerewashedwithwashsolutionand3.75x108nanoparticle‐bioconjugatesin40µlparticlebufferBalongwith10µLofcalibratorsolutionorsamplewasappliedtothesolid‐phasesurfacefollowedbya15‐minuteincubationat+36°C(650rpm).Aftersampleincubation,thewellswerewashedwithwash solution and time‐resolved fluorescencewasmeasuredfromtheassaywellsasdescribedinsection4.4.1.

The different versions of the cTnI assays developed during the study wereassigned a version number by which they are later referred to. The specificamountsofthesolid‐phaseantibodies,detectionparticles,andtheassignedassayversionnumbersarelistedinTable5.


44

Table5.AntibodyandparticleamountsofthedifferentcTnIassayversions.

Assayversion Solid‐phase Detection

StudyII(Assaysutilizingdifferentmolecularformsof9707)

1. Mab‐19C7+Mab‐9707(100+100ng/well) Mab‐8I7(3.75x108particles/well)

2. Mab‐19C7+F(ab')2‐9707(100+66ng/well)*

3. Mab‐19C7+Fab‐9707(25+16.5ng/well)

4. Mab‐19C7+cFab‐9707(25+16.5ng/well)

StudyIII(AssayutilizingcFabs)

5. Mab‐19C7+cFab‐9707(25+16.5ng/well) cFab‐11N11(3.75x108particles/well)

StudyIV(Assaysutilizingdifferentcommercialantibodies)

6. Mab‐19C7+Fab‐9707(50+33ng/well) Mab‐8I7(3.75x108particles/well)

7. Mab‐19C7+Fab‐9707(50+33ng/well) Mab‐625(3.75x108particles/well)

8. Mab‐19C7+Fab‐MF4(50+33ng/well) Mab‐8I7(3.75x108particles/well)

9. Mab‐19C7+Fab‐MF4(50+33ng/well) Mab‐625(3.75x108particles/well)

*(Järvenpääetal.,2012)

4.5 Assayevaluations

4.5.1 Methodologicalevaluation(I,III,IV)Assay sensitivities were assessed from the dose‐response curves and werecalculated either as the concentration deviating three standard deviations frombackground signal (analytical detection limit; II, IV), orby calculating the assaylimitofblank (LoB; I, III), limitofdetection (LoD; III), and limitofquantitation(LoQ; III) according to the Clinical and Laboratory Standards Institute (CLSI)GuidelinesEP17‐A2(ClinicalandLaboratoryStandards Institute,2012).Within‐run and within‐laboratory imprecisions were assessed according to the CLSIGuidelineEP5‐A2(III)(ClinicalandLaboratoryStandardsInstitute,2004).

Assayrecoveriesweredeterminedfromheparinplasmasamplesfortifiedwithanativetroponincomplexsothattheaddedvolumedidnotexceed5%ofthetotalsamplevolume(III).Thedeterminationofassaycross‐reactivitieswasexecutedin a similarmanner, but by using serum samples and skTnI (IV). The recoverypercentageswere calculated as the increase in cTnI concentration compared totheexpectedincrease,andthepercentualcross‐reactivitieswerecalculatedastheobserved skTnI‐signal converted to cTnI concentration and compared to theadded skTnI concentration. Assay linearity was assessed by diluting patientsamples containing known amounts of cTnI with a plasma pool with a signalbelowtheassayLoD(III).


45

4.5.2 Methodcomparisons(I,III)Samples described in section 4.1were used formethodological comparisons tocommercial assays. For D‐dimer the comparison assay was quantitativeimmunoturbidimetriclatexagglutinationSTALiatestD‐dimer‐assay(DiagnosticaStago, NJ, United States) (Waser et al., 2005). For themethod comparison, thedevelopedD‐dimerassaywasperformedwitha15‐minute incubation(I).Assayversion 5 for cTnI utilizing chimeric antibody fragments was compared toSiemensADVIACentaur®TnI‐Ultra™assay(Casalsetal.,2007)(III).

4.5.3 Assayinterferences(II–IV)Matrix‐related interferences were evaluated by measuring the apparent cTnIconcentrationsofheparinplasma samplesobtained from32apparentlyhealthyvolunteerswithassayversions1‒4,whereantibody9707wasmodified(II).TheeffectofusingcFabsasboththesolid‐phaseanddetectionantibodieswasstudiedwiththeassayversion5bymeasuringthecTnIvaluesof39samplescontainingknown amounts of triglycerides, bilirubin, RFs, or HAMA, and the results werethencomparedtothoseofthepreviouslypublishedversion2ofthesameassay(Järvenpääetal.,2012)(III).Theeffectofantibodyselectiononassayspecificitywas evaluated by comparing the calculated cross‐reactivity values of fourdifferentassayversions(6‒9)employingdifferentassayantibodies,soldascTnI‐specific, without changing the assay epitope specificity determined by peptidemapping.Also,aclinicalsamplepanelwasmeasuredwiththefourdifferentassayversions. Clinical samples that were observed to have a deviating assay resultwith the assay version 6were furthermeasuredwith an unpublished researchassayforskTnI(IV).

4.5.4 Statisticalanalyses(I–IV)Statistical analyses were conducted using SAS Enterprise guide 3.0 and SASsoftwareversion9.1(SASInstitute;NC,USA)(II)andIBMSPSSstatistics21(SPSSInc.,Chicago,IL)(I,III,IV).Thetwo‐tailedP‐values<0.05wereconsideredtobestatistically significant. Correlations between the different assays were studiedwith Spearman’s rank correlation (I–IV). In themethod comparison, Passing &BablokorweightedDemingregressionparameterswerecalculatedwithAnalyze‐it software (versions 2.30 (I,III) or 3.71.1 (IV), Analyze‐it Software Ltd., Leeds,UnitedKingdom).LinearregressionparameterswerecalculatedwithOriginDataAnalysis and Graphing Software (version 8.0, OriginLab Corporation,Northampton,MA)(II).TheagreementbetweendifferentassayswasassessedbyBland‐Altmananalyses(I,III)(BlandandAltman,1986).NonparametricKruskal‐WallisandMann‐WhitneyU‐testswereusedtocomparethedifferencesbetweenthe four different captures and between continuous variables of two differentcaptures(nongaussiandistribution)(II).

SummaryofResults

46

5 SUMMARYOFRESULTS

5.1 Nanoparticlesasuniversallabels(I)

5.1.1 AdjustmentofassaydynamicrangeTostudywhetherEunanoparticlescanuniversallybeapplied inheterogeneousimmunoassaysrequiringspeedandwidemeasuringrange,differentapproachesto extending immunoassay dynamic range were applied by using D‐dimer andcTnI asmodel analytes. Theultimate goalwas to develop assays thatwouldbeapplicable to POCT. Therefore, in order to simulate simple microfluidic chipprocessing where the sample and detection antibodies are combined prior totheir introduction to a solid‐phase surface, a 30‐second preincubation of thesampleandparticlesolutionwasanalyzed.

Thedesensitizationprocesswas analyzed so that theD‐dimerand cTnI‐specificantibodies(Mab‐FDP14,F(ab’)2‐8D3,Mab‐9707,orMab‐8I7)weretestedbothasthesolid‐phaseanddetectionantibodies.Theeffectofantibodyconfigurationandthe addition of free antibody on the upper limit of assay dynamic range,determinedvisually as thehighest calibrator concentration in the linearpart ofthecalibrationcurve,andLoBcanbeseeninTable6.

SummaryofResults

47

Table6.The calculated LoBs and dynamic ranges of the different assay versionswith 5‐minuteassay incubation. Modified from (I). The universally used clinical cut‐off level of D‐dimer is 500ng/mL(FEU)(Bounameauxetal.,2010).

LoB(ng/mL)DynamicRange(ng/mL)

D‐dimer F(ab')2‐8D3(solid‐phase)/MAb‐FDP14(detection) 0µg/mLfreeMab‐FDP14/F(ab')2‐8D3 0.249 ≤505µg/mLfreeF(ab')2‐8D3 0.463 ≤5015µg/mLfreeF(ab')2‐8D3 0.158 ≤2005µg/mLfreeMab‐FDP14 2.73 ≤100015µg/mLfreeMab‐FDP14 13.6 ≤10,000Mab‐FDP14(solid‐phase)/F(ab')2‐8D3(detection)0µg/mLfreeMab‐FDP14/F(ab')2‐8D32 1.17 ≤5001µg/mLfreeMab‐FDP14 2.75 ≤10,0005µg/mLfreeMab‐FDP14 478 ≤10,0005µg/mLfreeF(ab')2‐8D3 0.331 ≤100015µg/mLfreeF(ab')2‐8D3 0.603 ≤3000cTnIMab‐8I7(solid‐phase)/Mab‐9707(detection)0µg/mLfreeMab‐8I7/Mab‐9707 0.0100 ≤2510µg/mLfreeMab‐8I7 1.03 ≤505µg/mLfreeMab‐9707 0.0356 ≤10010µg/mLfreeMab‐9707 0.0808 ≤25025µg/mLfreeMab‐9707 0.0367 ≤1000Mab‐9707(solid‐phase)/Mab‐8I7(detection)0µg/mLfreeMab‐8I7/Mab‐9707 0.00300 ≤2510µg/mLfreeMab‐9707 1.06 ≤505µg/mLfreeMab‐8I7 0.00850 ≤5010µg/mLfreeMab‐8I7 0.0154 ≤10025µg/mLfreeMab‐8I7 0.290 ≤1000

WithD‐dimer,theadditionoffreeMab‐FDP14movedtheupperlimitofdynamicrangemorethantheadditionoffreeF(ab’)2‐8D3.Atbest,anadditionof15µg/mLof freeMab‐FDP14causeda200‐fold increase intheupper limitof thedynamicrangefrom50ng/mLto10,000ng/mL(F(ab’)2‐8D3/Mab‐FDP14‐configuration),and a simultaneous, nearly 55‐fold decrease in assay sensitivity (LoB 0.249ng/mLvs.13.6ng/mL).WithMab‐FDP14as the solid‐phaseandF(ab’)2‐8D3asthedetectionantibody,aslittleas1µg/mLfreeMab‐FDP14raisedtheupperlimitof thedynamicrange to10,000ng/mL.WithcTnI,bothantibodyconfigurationsreactedalmost identically to the freeantibodyaddition: thehighest raise in theupperlimitofdynamicrange,from25ng/mLto1000ng/mL,wasachievedwhen25µg/mLoffreedetectionantibodywasadded.

SummaryofResults

48

5.1.2 Performanceofa15‐minuteD‐dimerassayTo test the performance of theD‐dimer immunoassaywithMab‐FDP14 as thesolid‐phaseandF(ab')2‐8D3asthedetectionantibody(with1µg/mLfreeMab‐FDP14,Table6), theassaywasrunwith15‐minute incubationandtheresultswere compared to STA Liatest D‐dimer assay. Deming regression analysisyieldedaslope(95%confidenceintervals,CI)of0.0900(0.0700–0.110)anday‐interceptof‐7.79(‐17.9–2.29)ng/mL(Sy|x0.126ng/mL)(Figure9A).Themeanrelativedifference(95%limitsofagreement)withBland‐Altmanagreementwas170% (131%–209%) (Figure 9B). The Spearman correlation coefficient was0.906(P<0.001).Themeasuredmedianconcentrationsofthereferenceandthedevelopedassayswere(25th–75thpercentiles)1700ng/mL(575ng/mL;4485ng/mL,FEU)and111ng/mL(36.5ng/mL;354ng/mL),respectively.

Figure9.Methodcomparisonofnanoparticle‐basedimmunoassayforD‐dimer(15‐minuteformat)andSTALiatestD‐dimerassay.(A)Correlationbetweenthetwoassays(n=65,r=0.906).(B)Bland‐Altmananalysisofagreement.Themeandifferenceof the twoassays (170%) ispresentedwithasolid horizontal line and the dotted lines represent the 95% limits of agreement (131%–209%).Modifiedfrom(I).

5.2 AnalyticalperformanceofimmunoassaysforcardiactroponinI

5.2.1 Theeffectofsolid‐phaseantibody9707molecularform(II)To test the effect of an antibody molecular form on the measured cTnI‐levels,assay versions 1‒4 were compared. The utilization of different antibodyfragments did not significantly affect the assay signal level or its analyticalsensitivity(Figure10).Thecalculatedanalyticalsensitivities(background+3SD)were:assayversion1 (Mab)=0.90ng/L;assayversion2 (F(ab’)2)=0.91ng/L;assay version3 (Fab) = 0.69ng/L; and assay version4 (cFAb) = 0.41 ng/L.Allassayswerelinearupto10,000ng/L,andnohigh‐dosehookwasobservedevenwhen1,000,000ng/LcTnIwasmeasured.

101 102 103 104 105100

150

200

250BR

el.

diff

. (%

)

D-dimer (ng/mL, mean)100 101 102 103 104 105 106100

101

102

103

104A

New

ass

ay (

ng/m

L)

STA Liatest (ng/mL, FEU)

SummaryofResults

49

Figure10.Dose‐responsecurvesoftheassayversions1‒4utilizingdifferentformsofcTnIantibody9707.Theerrorbarsrepresentthestandarddeviationfromreplicatewells.Modifiedfrom(II).

5.2.2 Chimericrecombinantantibodyfragments(III)To possibly further decrease different matrix‐related interferences of the cTnIresearchassay,thedetectionantibody8I7wasreplacedwithacFab‐11N11inassayversion 5. The assay version 5was linear to 50,000 ng/L (R2=0.993). The assaysignalwasincreasingto500,000ng/L,andnohigh‐dosehookwasobserved.AssayLoBandLoDweredeterminedtobe1.35ng/Land3.30ng/L,respectively.Atypicaldose‐responsecurveoftheassay5canbeseeninFigure11.

Figure 11. A typical dose‐response curve of the assay version 5. The error bars represent thestandarddeviationfromreplicatewells.ThedottedlinesrepresentthecalculatedLoB(1.35ng/L)andLoD(3.30ng/L)oftheassay.Modifiedfrom(III).

101 102 103 104 105 106102

103

104

105

106

107

Assay version 1: y = 202.4x; R2 = 0.998




Flu

ores

cenc

e (c

ount

s)

cTnI (ng/L)

100 101 102 103 104 105 106 107101

102

103

104

105

106

107

y = 73.7x; R2 =0.993

Flu

ores

cenc

e (c

ount

s)

cTnI (ng/L)

SummaryofResults

50

Assayrecoverieswerestudiedbyspiking64plasmasamplesfromyounghealthyindividuals with 500 ng/L cTn complex. Assay recoveries, calculated as themeasuredcTnIfromthe500ng/LspikedcTnI,variedbetween70%and108%forsamplesfoundnegativefortroponinautoantibodies(average83%,median83%).Forthe5samplesfoundastroponinautoantibodypositive, theassayrecoveriesvaried between 31% and 78% (average 53%, median 57%). The samplebackgroundwasmeasuredwithoutaddedcTnI.Outofthe64samplesmeasured,4.7% (n=3) gave cTnI values above the assay LoD of 3.3 ng/L (5.74 ng/L, 7.20ng/L,and11.8ng/L).

Assaylinearitywasevaluatedusingsixpatientheparinplasmasamplesthatwerediluted1‒243‐foldwithapoolofnormalplasma.TheinitialcTnIconcentrationsof the samples were: 22,262 ng/L; 14,540 ng/L; 10,145 ng/L; 2327 ng/L; 979ng/L; and 718 ng/L. The assay showed good linearity (R2=0.929‒0.993)throughout themeasuredrange (2.2ng/L–22,261ng/L)as seen inFigure12A.Thewithin‐runandwithin‐laboratoryprecisionsdeterminedbyspikingapoolofcTnI negative plasma samples (UTU/BT) with endogenous cTnI obtained frompatient samples can be seen in Figure 12B. The measured mean cTnIconcentrations(twiceadayfor20days)were:29ng/L,37ng/L,and2819ng/Landthecorrespondingwithin‐runimprecisionswere8.5%,8.4%,and7.5%.Thewithin‐laboratory precisions were 13.7%, 16.4%, and 15.9%, respectively. TheassayLoQcalculatedasfunctionalsensitivitywithwithin‐laboratoryprecisionof20%was17ng/L.

Figure 12. A) Linearity of the assay version 5 assessed with serial dilution (1–243‐fold) of sixclinical heparin plasma samples. B) Within‐laboratory precision of the assay version 5, and thedetermination of assay LoQ (the lowest cTnI concentration measured with within‐laboratoryprecisionof20%).Modifiedfrom(III).

TocomparethecFab‐basedassaytoacommercialcTnIassay,265leftoverplasmasamples fromOuluUniversityHospitalweremeasuredwith theassayversion5andtheresultswerecomparedtothosemeasuredwithSiemensADVIACentaur®TnI‐Ultra™assay.Outofthemeasured265samples,17wereobserved<LoDwith

100 101 102 1030

10

20

30

40

50B

With

in-L

ab. P

rec.

(%

)

cTnI (ng/L)

LoQ(20%) = 17 ng/L

10-1

100

101

102

103

104

105 R2=0.955

R2=0.986

R2=0.993

A R2=0.934

R2=0.929

R2=0.948

cTnI

(ng

/L)

1/1 1/3 1/9 1/27 1/81 1/243Dilution of LiH-plasma

SummaryofResults

51

the assay version 5, and these were excluded from the method comparisonanalysis (6–300 ng/L; median, 20 ng/L). Passing & Pablok regression analysisyieldedaslopeof0.180(95%CI;0.170–0.190)anday‐interceptof1.94(95%CI;‐1.28‒3.91) ng/L (Figure 13A). The mean relative difference (95% limits ofagreement) with Bland‐Altman agreement was 136% (62.3%–202%) (Figure13B).TheSpearmancorrelationcoefficientwas0.965(P<0.001).Themeasuredmedian concentrations of the Siemens reference assay and the assay version 5were (25th–75th percentiles) 1480 ng/L (202 ng/L; 6640 ng/L) and 241 ng/L(37.1ng/L;1290ng/L),respectively.

Figure 13. Method comparison. (A) Correlation between the cTnI assay version 5 and SiemensADVIA Centaur® TnI‐Ultra™ reference assay (n=248 r=0.965). (B) Bland‐Altman analysis ofagreement between the two assays. Themean difference (136%) is presented with a horizontalsolid line.Thedashedsolid linesrepresent the95%limitsofagreement(62.3%–202%).Modifiedfrom(III).

5.2.3 DifferentcommercialcTnI‐specificantibodies(IV)To test cTnI specificity, four different antibodies from three differentmanufacturerswere tested in theresearchassay forcTnI,so thatnochanges intheantibodyconfigurationandepitopespecificityweremade.Theassayversions6‒9allusedMab‐19C7asthefirstsolid‐phaseantibody.Withthecombinationofeither Mab‐8I7 or Mab‐625 attached to intrinsically fluorescent nanoparticles,either Fab‐9707 or Fab‐MF4was used as the second solid‐phase antibody. TheimmunoassayswithMab‐8I7as thedetectionantibody(assayversions6and8)werelinearto~15,000ng/L,whereastheMab‐625‐basedassays(assayversions7and9)werelinearto~50,000ng/L.Thesignalswereincreasing,andnohigh‐dose hook was observed with 500,000 ng/L cTnI (Figure 14). The calculatedanalytical sensitivities were: assay version 6, 1.25 ng/L; assay version 7, 2.34ng/L;assayversion8,1.82ng/L;andassayversion9,3.84ng/L.

100 101 102 103 104 105 106

-200

-100

0

100

200B

Rel

. diff

. (%

)

cTnI (ng/L, mean)100 101 102 103 104 105 106100

101

102

103

104

105

106A

Ass

ay

vers

ion

5,

cT

nI (

ng/L

)

Advia Centaur, cTnI (ng/L)

SummaryofResults

52

Figure 14. Dose‐response curves of the four research assays versions utilizing differentcommercially available cTnI‐specific antibodies. The error bars represent the standard deviationfromreplicatewells.Modifiedfrom(IV).

5.3 CardiactroponinIimmunoassayspecificity

5.3.1 Theeffectofsolid‐phaseantibody9707molecularform(II)Theeffectofantibody9707molecular form(Mab,F(ab’)2,FabandcFab)on theapparent cTnI levels were assessed by measuring 32 plasma samples fromapparentlyhealthyreferencepopulationwith theassayversions1‒4.Minimum,maximum,median,andaveragevaluesmeasuredwithcTnIassaysexploitingthedifferentmolecular forms of antibody 9707 are presented inTable7. and as abox‐plot presentation in Figure 15. Values measured with assay version 1correlated poorly with assay versions 2, 3, and 4: r=0.220, 0.0300, and 0.010(P<0.001), respectively. The corresponding linear regression equations were:y=0.02166x+3.0937; y=0.002340x+2.7939; and y=0.0008870x+2.4106. TheSpearmancorrelationcoefficientsforassayversion2andversions3and4weresignificantlybetterr=0.770(assayversion3)andr=0.780(assayversion4).Thecorresponding linear regression equations were: y=0.6058x+0.7481 andy=0.6121x+0.3203, respectively. The correlation between the assay versions 3and4wasgood:r=0.970(y=0.9725x–0.3286).

101 102 103 104 105 106101

102

103

104

105

106

107





Flu

ores

cenc

e (c

ount

s)

cTnI (ng/L)

SummaryofResults

53

Table 7. The measured cTnI concentrations of 32 plasma samples from apparently healthyindividuals analyzedwith the cTnI assay versions utilizing differentmolecular forms of antibody9707.Modifiedfrom(II).

Assayversion Minimum (ng/L) Maximum (ng/L) Median (ng/L) Average(ng/L)

1 2.64 116 7.28 15.92 1.45 11.1 2.60 3.44

3 1.10 11.2 2.51 2.83

4 0.746 10.6 1.80 2.42

Statistical analysis of the cTnI values measured with the four different assayversionsshowedsignificantdifferencesbetween the four9707molecular forms(P<0.001, Kruskal‐Wallis test). Mann‐Whitney test also showed statisticaldifferencesbetweenassayversion1andtherestofthetestedassayversions2‒4(P<0.001). The difference between assay versions 2 and 3, as well as betweenversions3and4wasinsignificant:P=0.38andP=0.10,respectively.However,thedifference between assay versions 2 and 4 was observed to be statisticallysignificant:P=0.016(Mann‐Whitneytest).

Figure15.Box‐plotpresentationof themeasuredcTnIvalues (ng/L) from32apparentlyhealthyvolunteersmeasuredwithassayversionsutilizingdifferentmolecularformsof9707antibody.Thecrosses indicate theminimumandmaximumvalues; theboxes represent the25‒75thpercentiles;thewhiskersthe1‒99thpercentiles;thehorizontallinesrepresentthemedianandthesmallboxesrepresentthemean.Modifiedfrom(II).

5.3.2 Chimericrecombinantantibodyfragments(III)Theeffectofutilizing chimericantibody fragments (assayversion5)onmatrix‐related interferences was assessed by comparing measured cTnI levels fromsamples containing known amounts of triglycerides, bilirubin, RF, or HAMA tovaluesmeasuredwith a previously published assay version 2 (Järvenpää etal.,2012).TheresultsfromthetwoassayscanbeseeninTable8.

1 2 3 4

100

101

102

103

Assay version

cTnI (

ng/L

)

SummaryofResults

54

Table 8. Comparison of plasma and serum samples containing known amounts of triglycerides,bilirubin, RFs, or HAMA measured with assay version 2 and assay version 5 employing cFab‐fragments.Modifiedfrom(III).

Assayversion2(LoD=2.0ng/L)

Assayversion5(LoD=3.30ng/L)

Concentrationofpossiblyinterfering

substance

Interferingsubstance measuredcTnIconcentration(ng/L)

measuredcTnIconcentration(ng/L)

5.73g/L Triglycerides <LoD <LoD18.16g/L Triglycerides <LoD <LoD10.55g/L Triglycerides <LoD <LoD10.95g/L Triglycerides <LoD <LoD48.19g/L Triglycerides 4.09 4.3913.42g/L Triglycerides <LoD <LoD11.02g/L Triglycerides <LoD <LoD11.59g/L Triglycerides 6.30 3.2710.57g/L Triglycerides <LoD <LoD8.88g/L Triglycerides <LoD <LoD7.53g/L Triglycerides 2.23 <LoD10.41g/L Triglycerides 3.31 <LoD0.002g/L Bilirubin 13.6 5.340.002g/L Bilirubin <LoD <LoD0.05g/L Bilirubin <LoD <LoD0.05g/L Bilirubin <LoD <LoD0.11g/L Bilirubin 20.6 10.10.10g/L Bilirubin 10.4 <LoD0.17g/L Bilirubin 4.75 <LoD0.14g/L Bilirubin 7.39 <LoD0.18g/L Bilirubin 7.64 3.920.32g/L Bilirubin 3.64 <LoD0.32g/L Bilirubin 7.01 <LoD0.14g/L Bilirubin 26.7 10.70.10IU/L RF 6.3 <LoD0.10IU/L RF <LoD <LoD0.03IU/L RF <LoD <LoD0.34IU/L RF <LoD <LoD0.34IU/L RF 3.81 <LoD0.03IU/L RF 2.12 <LoD1.06IU/L RF 7.43 <LoD0.79IU/L RF <LoD <LoD3.27IU/L RF 241 <LoD1.00IU/L RF 3.04 <LoD1.29IU/L RF <LoD <LoD0.70IU/L RF 2.92 <LoD0.30g/L HAMA <LoD <LoD0.14g/L HAMA <LoD <LoD0.18g/L HAMA <LoD <LoD

IU,internationalunit

SummaryofResults

55

Inassayversion5,15.4%(6/39)ofthesamplesgavecTnIconcentrations>LoD,when the corresponding value for assay version 2 was 51.3% (20/39). ByremovingtheFc‐partfromtwooutofthreeantibodies,interferencefromRFwastotallyeliminatedinassayversion5:thehighestvaluebeing241ng/Lobservedwithassayversion2.Outofthe12samplescontainingbilirubin,33.3%(4)gaveameasurablecTnIvalue>LoDwithassayversion5,whenthecorrespondingvaluewas75.0%(9)withassayversion2.

5.3.3 DifferentcommercialcTnI‐specificantibodies(IV)Four different antibodies originally from three different manufacturers weretested as solid‐phase – detector pairs in four different combinations (assayversions6‒9)sothattheepitopespecificityoftheassaywasretained.Atotalof101 samples were analyzed with all of the four different assay versions. TheDeming regressionanalyses for all sixdifferent comparisonswith95%CIs, andthe Spearman correlation coefficients (P<0.001) can be seen inFigure16. Themeasuredmedianconcentrationsoftheassayswere(25th–75thpercentiles)583ng/L(76.0ng/L,3290ng/L,assay6);360ng/L(55.0ng/L,2260ng/L,assay7);390ng/L(46.0ng/L,2570ng/L,assay8);and268ng/L(45.0ng/L,2240ng/L,assay9).

SummaryofResults

56

Figure16.Clinical samples (n=101) testedwith fourdifferentassayversions6‒9.Filled symbolsrepresent samples that were observed to give unexpectedly high signals with assay version 6.Dashedlinesrepresentthelinesofidentity.Modifiedfrom(IV).

Out of the 101 heparin plasma samples tested, five were observed to giveunexpectedlyhighvalueswithassayversion6,butwerecorrectedwiththeotherversions(filledsymbolsinFigure16).Whenthosesamplesweretestedwithaninvestigational research immunoassay for skTnI (unpublished), they were allobservedtocontainskTnI,thevaluesrangingfrom5500ng/Lto702,000ng/L.

y = -3.141 + 0.742 xslope 95% CI: 0.669 ̶ 0.819y-int. 95% CI: -5.24 ̶ -1.04r = 0.942

y = -1.97 + 0.729 xslope 95% CI: 0.657 ̶ 0.801y-int. 95% CI: -4.14 ̶ 0.188r = 0.937

y = -4.01 + 0.698 xslope 95% CI: 0.613 ̶ 0.783y-int. 95% CI: -7.27 ̶ -0.760r = 0.929

y = 1.03 + 0.982 xslope 95% CI: 0.922 ̶ 1.04y-int. 95% CI: -0.503 ̶ 2.56r = 0.989

y = -1.59 + 0.923 xslope 95% CI: 0.864 ̶ 0.982y-int. 95% CI: -3.56 ̶ 0.468r = 0.991

y = - 2.65 + 0.940 xslope 95% CI: 0.879 ̶ 1.00y-int. 95% CI: -4.58 ̶ 0.730r = 0.993

100 101 102 103 104 105100

101

102

103

104

105A

ssa

y 7

(n

g/L

)

Assay 6 (ng/L)100 101 102 103 104 105100

101

102

103

104

105

Ass

ay 8

(ng

/L)

Assay 6 (ng/L)

100 101 102 103 104 105100

101

102

103

104

105

Ass

ay 9

(ng

/L)

Assay 6 (ng/L)100 101 102 103 104 105100

101

102

103

104

105

Ass

ay

8 (

ng

/L)

Assay 7 (ng/L)

100 101 102 103 104 105100

101

102

103

104

105

Ass

ay

9 (

ng

/L)

Assay 7 (ng/L)100 101 102 103 104 105100

101

102

103

104

105

Ass

ay 9

(ng

/L)

Assay 8 (ng/L)

SummaryofResults

57

WhenFab‐9707andMab‐8I7wereusedtogetherasasolid‐phase–detectionpairinassayversion6,and5000,50,000,and500,000ng/LskTnIwereadded,afalsecTnI signal up to 7867 ng/L wasmeasured. ChangingMab‐8I7 to Mab‐625, orFab‐9707toFab‐MF4decreasedthemeasuredfalsecTnIsignaltoaboutonetenth(616–727ng/L).ThemeasuredfalselyelevatedcTnIsignalsofeachoftheassayversionscausedbyskTnIadditionscanbeseeninTable9.

Table9. Different spiked skTnI amountsmeasured as falsely elevated cTnI with different assayversions utilizing antibodies originally from different commercial manufacturers. Modified from(IV).

Assayversion: 6 7 8 9

skTnI Measuredasng/LcTnI5000ng/L 17 10 7 10

50,000ng/L 286 77 68 65

500,000ng/L 7867 727 652 616

Thecalculatedpercentualcross‐reactivitiesfordifferentassayversionswhen50or500ng/LcTnIwerecombinedwith50,000or500,000ng/LskTnIcanbeseenin Table 10. Assay version 6 showed the highest cross‐reactivity with up to1.36%.Allotherversionshadcross‐reactivitiesrangingfrom0.04%to0.21%.

Table10.Thecalculatedcross‐reactivitieswithdifferentassayversions,when50,000or500,000ng/LskTnIweretestedwith50and500ng/LcTnI.Modifiedfrom(IV).

Assayversion 6 7 8 9

cTnI skTnI Cross‐reactivity

50ng/L 50,000ng/L 0.42% 0.17% 0.12% 0.12%

500,000ng/L 1.36% 0.19% 0.12% 0.13%

500ng/L 50,000ng/L 0.44% 0.17% 0.04% 0.15%

500,000ng/L 1.27% 0.21% 0.13% 0.15%

Discussion

58

6 DISCUSSION

Aslongasantibodieshavebeenusedinclinicalassays,differentinterferenceshavebeenamatterofdiscussion.Theestimationisthatheterophilicinterferenceisthereason forone falseassay result forevery2000assaysperformed(LevinsonandMiller,2002).Therefore,interferencesare,fromtimetotime,alsoobservedincTndiagnostics. Falsely elevated cTn levels emerging from heterophilic‐typeinterferenceshavebeenreportedduringthepast10yearsforbothcTnTandcTnIinmultipleassaysystems (Lumetal., 2006;Biondaetal.,2007;Zhuetal.,2008;Shayanfar etal., 2008; Pernet etal., 2008; Ghali etal., 2012; Lippi etal., 2013a).Negative interference from cTn‐specific autoantibodies is also becomingincreasinglyrecognized,althoughitwasalreadydiscussedinthe1990s(Bohneretal.,1996).

6.1 Immunoassaysandtheirperformance

6.1.1 Nanoparticlesaslabels(I)Eu nanoparticles have proven their utility in immunoassay approaches, whereassay sensitivity is highly valued (Järvenpää et al., 2012; Valanne et al., 2005;Soukka et al., 2003). In original publication I, increasing the amount of freeantibody resulted in less sensitive assay in a majority of the different assayversions.Thiscanbeexplainedbytheassumptionthatincreasinglyfewantigenswereabletobindbothsolid‐phaseanddetectionantibodies,whentheamountoffreeantibodywasincreased.Theresultsalsoclearlyemphasizetheimportanceofcarefullystudyingtheconfigurationofantibodiesintwo‐siteimmunoassays:withnoadditionoffreeD‐dimerantibodies,almostafive‐folddifferenceinassayLoBandaten‐folddifferenceintheupperlimitofdynamicrangewereobserved,withchanging the antibody configuration. To further emphasize the importance ofantibodyorientationand theadditionof freeantibodies,anumberof the testedassayversionswouldhavebeenclinicallyuselesswithdynamicrangeswellbelowtheclinicallyinterestingareaof500ng/mL(FEU).ForcTnI,thedifferenceswereclearly less significant. Specific reports focusing on the extension of assaydynamic ranges are very limited. These include the use of specific molecularspacers and plasmonic structures enabling an increase of up to 8 orders ofmagnitude inassaydynamicranges,whencomparedtotraditionalassays(Zhouetal., 2012). Also the application ofmultiple antibodies possessing a variety ofaffinities (Ohmuraetal., 2003) and theuse of immunosensorswith fluorescentlabels (Renard and Bedouelle, 2004) have been introduced. Traditionally, theextension of immunoassay dynamic range is approached with the desire toachieveassensitiveanassayaspossible,withtheneedtouseonlysmallamountsofthesample(Toddetal.,2007).Extensionofassaydynamicrangebydilutingthesampleordecreasingthesamplevolumewasnotusedhereduetotheassaybeingamodel system for a commercial assay formatwithpredetermined sample andbufferratios.Whentime‐resolvedfluorescence isused, theapplicationofexcess

Discussion

59

functional binders usually offers an opportunity for wide dynamic ranges(Lövgren etal., 1984), aswell asmeans to shorten the incubation time (Ekins,1960).

InthecriticalcareunitsandEDs,theaverageTATexpectedbythephysiciansis5‒15minutes(Harvey,1999).Duetothehighspecificactivityofthelanthanide‐dopednanoparticles,itwaspossibletodevelopimmunoassayswithshort5‐and15‐minuteincubations.Therefore,itcanbestatedthatnanoparticlescanbeusedasuniversallabels,whendevelopingnovelimmunoassaysforPOCTapplications.

6.1.2 ImmunoassayversionsforcardiactroponinI(II‒IV)The current official recommendation of Clinical Chemistry and LaboratoryMedicine for cTnI assays is that the utilized antibodies should target the stablemidfragmentoftheanalyte(Panteghinietal.,2001).Ithasbeenreported,though,thatapproximately10%ofthepatientswithchestpainhavecTnIautoantibodiesin their circulation, thus possibly resulting in falsely decreased cTnI values(Savukoski et al., 2014). This is caused by the fact that the circulating cTnautoantibodiesaremainlyboundtothestablemidfragmentofcTnI(Erikssonetal.,2005b;Savukoskietal.,2012;Savukoskietal.,2013).OriginalpublicationsII‒IVwereallbasedonanassayconfiguration,whereallthethreeassayantibodieswereselectedonthebasisof theirability tocircumvent the inhibitingeffectsofthecirculatingcTnautoantibodies(Erikssonetal.,2005b).However,whenassayrecoverieswerestudiedinpublicationIII,therecoveryvalueswerelowerforcTnautoantibody‐positivesamples:themeanandmedianvalueswere53%and57%,whenthecorrespondingvaluesforcTnautoantibody‐negativesampleswere83%and83%,respectively.Thisdifferencebetweentherecoveriesfromautoantibody‐positive and negative samples can partly be explained with the avidity of thedetection particles. By attachingmultiple detection antibodies to the surface ofthe nanoparticles, we are increasing the binding site density through avidity,whichhasbeenshowntoenhancetheinherentaffinityoftheantibody(Soukkaetal., 2001a). Thus, the high density of the antibody renders the assay morevulnerable to all matrix‐related analytical interferences, including those ofcirculatingautoantibodies.IthasalsobeenshownthatinterferencefromcardiacautoantibodiesmaynotbetotallypreventedevenbyusingacTnIassaydesignedto avoid the central parts of the cTnI molecule most affected by autoantibodyinterference(Erikssonetal.,2005b).

Themolecularformofthe9707antibodydidnotaffectthesensitivityofthecTnIassay.InoriginalpublicationII,whereMab,F(ab’)2‐,Fab‐,andcFab‐fragmentsofantibody9707weretestedasapartofdifferentversions(1‒4)ofacTnIresearchassay,itwaspossibletoreducetheantibodyamountsofthe9707antibodyfrom0.63pmol(assayversion1)to0.31pmol(assayversions3and4).Also,withtheemployment of recombinant 9707 fragments, the amount of the second solid‐phase antibody Mab‐19C7 was decreased to one fourth. These reductions inantibodyamountscanmostprobablybeexplainedbymoreorientedsolid‐phase

Discussion

60

surfacesenablingalargernumberofavailableantibodybindingsitesthancanbeachievedwithtraditionallybiotinylatedMabandF(ab’)2‐fragments(Pelusoetal.,2003). To highlight the nonexistent effect of antibodymolecular form on assaysensitivityevenmore,thecalculatedLoD(3.30ng/L)ofassayversion5(III)wasinthesamerangeaswiththepreviouslypublishedassay2(2.0ng/L)(Järvenpääetal.,2012).

In original publication III, an assay version 5 utilizing cFabs was introduced.Probablyduetothefactthatthenormalsampleswereobtainedfromyoungandapparently healthy individuals who are unlikely to present with increased cTnvaluescommonlyfoundinelderlypeople(Vengeetal.,2003;Eggersetal.,2013),only 3 (4.7%) of the 64 measured samples gave signals >LoD. Altogether, themeasured cTnI concentrationsof assayversion5were substantially lower thanthosemeasuredwith thecommercial referenceassay(SiemensADVIACentaur®TnI‐Ultra™):themeandifferencewas136%.Furthermore,outofthe265patientsamples measured with the assay version 5, 17 gave cTnI values <LoD. Thereference assay utilizes remarkably different antibody epitopes than the cTnIresearchassay.TheSiemensreferenceassayemploysthreeMabswithantibodiesrecognizingaminoacidresidues27–40,41–49,and87–91andanLoBof6ng/L(Casalsetal.,2007),whilethecTnIresearchassayutilizesantibodieswithepitopespecificitiestoaminoacidresidues41–49,160–179,and190–196.Thisdifferencein the utilized antibodies may explain the significantly different cTnI valuesmeasuredwiththetwoassays.Ithasbeenreportedthatbothantibodyaffinitiesaswell as the utilized epitopes can affect themeasured signals evenwhen theassays are tested with the same calibration material (Savukoski et al., 2012).Therefore,thecombinationofusedcalibrationmaterialanddifferentantibodiesisthemost probable reason for the assaybias and for the17 sampleswith assayvalues<LoD.Another reasonmaybe the fact thatby replacing twoof the threeoriginal antibodies with recombinant cFabs, the interferences caused byheterophilic‐type interferencesweredecreased.Thishypothesis issupportedbythefactthatthe17samplesforthemostpartrepresentedlowendvalueswiththeSiemensreferenceassay(6–300ng/L,median20ng/L).

The basic quality requirements for the POCT of cTnI do not differ from otherassays:ideally,POCTassaysforcTnIshouldmeetthecurrentrequirementof10%within‐laboratory precision recommended for cTnI‐assays at MI decision limit(Wu et al., 1999). Currently, there are nine POCT devices available for thedetection of cTnI and cTnT that demonstrate considerable differences in theirabilitiestorule‐inandrule‐outMI(Palamalaietal.,2013;AmundsonandApple,2014).ThecFab‐basedassayversion5wasconductedmanuallywitha15‐minuteassay incubation. Therefore, the assay signal was measured at a highly kineticstage(13–41%ofmaximumsignal), thussignificantlyhampering theevaluationofassayLoQandprecision.Theassaywasnotabletomeetthecurrentprecisionrequirements, but applying the assay as a part of an automated assay systemwouldmostprobablyfacilitatebetterassayprecisionandenablehittingthe10%LoQgoal.To reduce theTAT, assays forPOCTshouldalsobeable tousewhole

Discussion

61

blood. However, in this study, citrate and heparin plasma as well as serumsampleswere used, since all the assayswere performed in a simplemicrotiterwellformat.OnlybyincludingtheassaysinafullyfunctionalPOCT‐system,wouldit be possible to fully evaluate the POCT applicability of the developednanoparticle‐assistedimmunoassays.

6.2 Matrix‐relatedinterferencesMatrix‐related interferences were studied in original publications II‒IV.Traditionally, interferences incTnIandcTnT‐testinghavenotbeenconsidered tobe a significant problem, but as the assay LoDs are decreasing with the newgenerationofhighsensitivityassays,theneedtore‐evaluatetheissuehasraiseditshead(MorrowandAntman,2009).Theexact incidenceof interferencesoccurringincTnassays isunknown,but it isexpectedtobehigherthanpreviouslythought(Zaidi and Cowell, 2010). In a recent study, it was observed that 1.01% of allmeasuredcTnIvalueswere falselyelevated (Lietal.,2014). In studies II and III,mouse Mabs, mouse Fabs, and cFabs were utilized. As the assay interferencesdecreasewiththeFc‐regionremovalandantibodychimerization, thenextstep inreducing thematrix‐related interferenceswouldbe thehumanizationof theusedcFabs by replacing the remaining mouse framework parts in the VH and VLsequenceswiththecorrespondingpartsofhumanorigin(Gonzales,2003).

As stated previously, the cTnI research assay configuration (II‒IV) is based onthree antibodies: two solid‐phase antibodies and one detection antibody. Thistypeof three‐site immunoassay approachhavebeen shown tobeover twice asprone to heterophilic antibody interference as traditional two‐site assayapproaches(Zhuetal.,2008).OneofthecTnIassaysolid‐phaseMabsrepresentssubclassIgG2(19C7),andtheremaining(9707,8I7,and625)belongtosubclassIgG1, which have been found to be highly susceptible to heterophilic‐typeinterferences (Bjerner et al., 2005). Therefore, the interferences observed inoriginal publication II, especially with assay version 1, are probably stemmingfrom heterophilic interferences. As with other analytes, heterophilic‐typeinterferencesareextremelydifficult topredictwithcTnI. Interferencesbetween0.1%and3.1%havebeenreported inanormalpopulation,andan interferencerateashighas50%wasobservedinapopulationwithcertaintypeofpneumonia(Yeo etal., 2000;Uettwiller‐Geiger etal., 2002;Kim etal., 2002; Fleming etal.,2002; García‐Mancebo et al., 2005). Other causes than heterophilic and similarantibodies for possible spuriously increased cTnI values cannot be ignored, asfalselyelevatedcTnvalueshavealsobeenconnectedtofibrinandtothepresenceofmicroparticles(Lietal.,2014).

The normal range values measured with the different assay versions (II, III)utilizingantibodyfragmentswereinthesamerangeasreportedpreviouslywithdifferentresearchversionsofhighsensitivityassays.Vengeetal.(2009)reportedthataprototypeofBeckmanCoulterAccesshighsensitivitycTnIassaywasabletomeasure normal range cTnI values between 1.1‒7.9 ng/L (median 3.2 ng/L).

Discussion

62

Instead,Wuetal.(2009)havereportedamedianvalueof1.72ng/LforSingulexhigh sensitivity assay. In publication II, assay versions 2, 3, and 4 hadmediannormal range values of 2.60 ng/L, 2.51 ng/L, and 1.80 ng/L, respectively. Thecorresponding minimum/maximum values were: assay version 2, 1.45/11.1;assayversion3,1.10/11.2ng/L;andassayversion4,0.746/10.6ng/L.EventhenormalvaluesabovetheassayLoDmeasuredwiththeversion5oftheresearchassay (5.74 ng/L, 7.20 ng/L, and 11.8 ng/L) fall within the same range aspublishedforotherresearchassays(III).Nevertheless,itmustbenotedthatdueto differences in the used calibrators and antibodies, the measured valuesbetween different assays are not directly comparable from assay to assay(Christensonetal.,2006).

When assay versions 2 and 5 were compared with samples containing knownamounts of triglycerides, bilirubin, RF, or HAMA, clear reductions in the cTnIlevelsofmeasured sampleswereobserved (III).However,nocomparisoncTnI‐valuewasavailableforthesesamples,soitcannotbeconcludedthatthesampleswere from apparently healthy persons with a low expectancy of measurableamounts of cTnI. The only difference between assay versions 2 and 5 was theremoval of the Fc‐portion in two out of three antibodies used, and antibodychimerization. Thus, it can be concluded that themajority of differences in theobserved cTnI values were stemming from the utilization of recombinantantibodyfragments.Valuesverylittle>LoDwiththeassayversion2and<LoDoftheassayversion5maybeduetothesmalldifferencesinassayLoDsratherthanthereductionofassayinterference.Thegreatestobserveddifferencewasashighas241ng/Lvs.<LoD.Thedifferencesshouldnotbeofthismagnitudewhenusingtwodifferentversionsorthesameassay,unlesstheotherismoresusceptibletomatrix‐related interferences. When possible interferences are found, therecommendation is that recovery and linearity studies be conducted to confirmthat true interference is being detected (Sturgeon and Viljoen, 2011; Ismail,2007).However,thevolumeandnumberofthesamplesobtainedfromProMedDxwaslimited,sothatonlycTnIvalueswiththetwoversionsoftheresearchassayweremeasured.Thelownumber(n=3)ofHAMAcontainingsamplesresultedinnone of the HAMA containing samples having a cTnI concentration >LoD withassay version 2. Thismeans that the HAMAs present in the three sampleswasmostprobablyasubtypethatdoesnotaffecttheantibodiesinassayversions2or5.Asexpected,theRFinterferencewastotallyeliminatedinassayversion5.Alsothe samples containing bilirubin showed decreases in themeasured cTnI. Thiswas unexpected, since false negative cTnI values have usually been associatedwith samples containing bilirubin (ver Elst et al., 1999; Dasgupta et al., 2001).However, spuriously increased cTnI values have been reported in samplescontaining free hemoglobin the generation of which is often accompanied byincreased bilirubin values (Hawkins, 2003;Masimasi andMeans, 2005). On theother hand, since no reference cTnI value was obtained for the interferencesamples, it cannotbe concluded for certaintywhether themeasured cTnI valuewithassayversion2 in triglycerideandbilirubincontainingsamplesrepresentsfalseelevations.

Discussion

63

AlthoughhighsensitivitycTnassaysoffersuperiorsensitivityandprecisionwhencomparedtocontemporaryassays(Christetal.,2011;WuandChristenson,2013;Korley and Jaffe, 2013), a problemwith lowered assay specificity has emerged.Actually, high sensitivity cTn assays are able to detect positive troponin values(>99thpercentile)inavarietyofdifferentnon‐ACS‐relatedconditionsincludingawide range of non‐ischemic clinical conditions, aswell as in sepsis and even inextremeexertion(Hammetal.,2002;Agewalletal.,2011).Melansonetal.(2008)reported thatwhen an old cTn assay fromSiemensHealthcareDiagnosticswasreplacedwith amore sensitive TnI‐Ultra in 2007, a doubling of the amount ofpositivecTnresultsinthecollectedsampleswasobserved,althoughnochangeintheactualnumberoffinaldiagnoseswerefound.

Assay cross‐reactivity issueswith cTns have notwidely been addressed sincethe cross‐reactivity problems were observed with the first‐generation cTnTassay. The affected assay employed a detection antibody that had 12% cross‐reactivity with skeletal troponin T (skTnT) (Katus et al., 1992; Gaze andCollinson,2008).Thecross‐reactivitywaseliminated in thesecond‐generationassay(Ricchiutietal.,1998),butithasre‐emergedasaconcernduringthepastfew years (Jaffe et al., 2011). Cross‐reactivity studies are mandatory forcommercial assays, but not all cross‐reactivity values are openly reported.According to instructions for use ‐leaflets, high sensitivity cTnI assays havevarying cross‐reactivities determined with 1,000,000 ng/L skTnI: AbbottArchitect0.07%andBeckmanAccess20.034%.Thismeansthat, theoretically,50,000 ng/L circulating skTnI would give rise to a 35 ng/L apparent cTnIconcentration with the Abbott Architect assay, which would exceed thedetermined assay 99th percentiles (women 11.4 ng/L,men 27 ng/L, and both19.3ng/L)(Krintusetal.,2014).Wuetal. (2009)havereported thatnocross‐reactivitywasobservedforSingulexErennahighsensitivitycTnIassay.Itmustbe noted, however, that skTnI was assayed over a range of 0.1–100 ng/L.Normal population are reported to have circulating skTnI concentrations ofapproximately 5500 (±5200) ng/L measured before exercise and 89,500(±71,400) ng/L after anaerobic exercise (Chapman etal., 2013). Ultimately ashigh as 990,000 ng/L skTnI has been measured in a patient with aninflammatory muscle disease polymyositis (Kiely, 2000). Therefore, with theincreasingly sensitive troponin assays, even small increases in skTnI maypotentiallybeasourceoffalselyincreasedtroponinvalues.

InstudyIV,thecombinationofantibodies9707and8I7assolid‐phase–detectionpair(assayversion6)causedclearlythestrongestcross‐reactivitytoskTnI(upto1.36%). By replacing either 9707 or 8I7 with MF4 or 625 antibody, markedlylowered the cross‐reactivity values. As it has become obvious that increasedfalselyelevatedcTnvaluescanbemeasuredwithoutMI(Lippietal.,2013a),moreattention shouldbedirected to assay cross‐reactivities.Manymanufactures sellMabs as cardiac‐specific without any documentation from possible specificitystudies.Therefore,theselectionofassayantibodiesbecomescruciallyimportantinattempts todevelopthenextgenerationofhighsensitivityassays.Antibodies

Discussion

64

sold as cardiac‐specific can detect trace amounts of skTnI, as was observed inpublicationIV.Vylegzhaninaetal.(2013)recentlyreportedthatMab‐8I7has70%cross‐reactivitywith skTnI. Results from study IV support the observation, butalsohighlighttheimportanceoftheotherantibodiesusedintheassaysetup.Mab‐8I7 showedextensive cross‐reactivitywith skTnIonlywhenused togetherwithFab‐9707 (0.42%–1.36%, assay version 6). Assays employing only 8I7 or 9707hadcross‐reactivitiesbetween0.04%–0.21%(assayversions7and8).Also, theassayversion9thatemployedantibodieswithnoreportedcross‐reactivitywithskTnI (from HyTest Ltd.), had cross‐reactivities in the same range (0.12%–0.15%).Intheory,evena0.04%cross‐reactivity,couldcausefalselyelevatedcTnIvalues.Inpracticethough,onlyassayversion6withboth9707and8I7antibodieswasclearlyaffectedbyskTnIpresentinthepatientsamples.Nevertheless,sinceassay versions 7‒9 showed identical cross‐reactivities, the specificity of assayantibodies should not be ignored, when developing new generations of highsensitivityassays.

The specificity issues aroundhigh sensitivity cTnassays shouldnotbe ignored,since they are routinely being implemented in clinical use. The lowered assayspecificityhasledtodiscussionsaboutthevalidityofthe99thpercentilevaluesinclinicaldecisionmaking(MorrowandAntman,2009;KhalilianddeLemos,2014;Aakreetal.,2014;Simpsonetal.,2014).Appleetal.(2012b)havereportedthatup to 32‐fold differences can be found in the assay 99th percentiles, whendeterminedfor19cTnIandcTnTassayswiththesamereferencepopulationused.Thus, the question about theway normality should be determined has becomehighly topical. In the future, the determination of assay 99th percentiles will,therefore, probably be based on strict recommendations and on the extensivemedical examinations of the study participants (Koerbin etal., 2013). AlthoughthecurrentrecommendationofMIdiagnosisemphasizestheimportanceofariseand/or fall in themeasured cTnvalue (Thygesenetal., 2012), an analyte valueabovethedetermineddiagnosticthresholdmay,insomecases,betheonlybasisfor diagnosing an ischemic cardiac disease (Eggers et al., 2012; Lippi et al.,2013b). In theworst casescenario, a falsepositive,or falsenegative, cTn‐resultmaythereforeresultinawrongclinicaldecision.

As there is no single blocking agent that could remove all heterophilic‐typeinterferences, andmost commercial immunoassays for cTnI are currently usingintactmonoclonalorpolyclonalantibodies(Tate,2008),employingrecombinantantibody fragments ‒ chimeric or humanized in particular ‒ should be highlyencouragedinthefuturegenerationsofhighsensitivityassaysforcTnIandTnT.Additionally, since the trend has been toward increasing sensitivity, and a cTnassayistoberegardedasahighsensitivityassayonlywhen>50%(ideally95%)ofpeopleinanormalpopulationhavedetectablecTnIvalues(Appleetal.,2012a),it shouldbecarefully studied,whether thesehighsensitivityassaysareactuallymeasuringcTn,andnotsomethingmatrix‐related.

Conclusions

65

7 CONCLUSIONS

IncreasinglysensitiveassaysarebeingdevelopedforcTnI,whichhasresultedinsignificantly lowered assay specificities. Therefore, the selection of assayantibodies plays an important role in the development of assays with highspecificity to cTnI. In addition, the employment of recombinant antibodyfragments,chimericfragmentsor,ultimately,humanizedantibodiespossesshighpotential for developing immunoassays minimally prone to matrix‐relatedinterferencespossiblypresentinallbloodsamples.

Themainconclusionsbasedontheoriginalpublicationsare:

I Theselectionofcorrectantibodyorientationaswellas theexploitationoffreesolid‐phaseordetectionantibodiesinthecapturesolutionenabledtheextension of the upper limit of assay dynamic range up to 200‐fold innanoparticle‐assisted immunoassays.Theextensionofdynamicrangewascarriedoutwithtwodifferentanalytesrequiringverydifferentsensitivitiesandassayranges.Therefore,itcanbeconcludedthatEu‐dopedintrinsicallyfluorescent nanoparticles are universally applicable to immunoassaysrequiringeitherhighsensitivitiesorwidedynamicrangesaroundaspecificpredeterminedanalytevalue.

II A possibly interference‐prone subclass IgG1 antibody was tested in aresearchassayforcTnIinfourdifferentmolecularforms:Mab,F(ab’)2,Fab,andcFab.UtilizingtheMabformcausedsignificantlyhigherobservedcTnIvalues in apparently healthy study subjects. Ultimately, the recombinantfragments showed the lowestmeasured cTnI values, calling for a broaderevaluationof theadvantagesofrecombinantantibodyfragments inassaysrequiringhighsensitivities.

III Highly sensitive assays for cTnI have emerged in the critical care testingscene during the past few years. Increasing the sensitivity has ultimatelymeantdecreasesinassayspecificities,thuscausingconcernsinestablishingandutilizingcorrectassay99thpercentiles.Anovel immunoassay forcTnIbasedonchimericFab‐fragmentsenabledtherapidmeasurementofcTnIaswell as significantly lowered the prevalence of spuriously elevated cTnIvalueswhen compared to a previous version of the assay. This highlightsthe results of study II, and calls for thorough and active measures inapplying chimeric antibody fragments when the next generations of highsensitivitycTnIassaysarebeingdeveloped.

IV Antibodies claimed as cTnI‐specific are widely sold by different antibodymanufacturers. However, these antibodies may cause falsely elevatedsignalsthroughcross‐reactivitywithskeletalTnI,especiallywhentheyareemployedinnanoparticle‐assistedimmunoassays,whereanavidity‐relatedenhancement of the assay signal may induce even low‐affinity matrix‐

Conclusions

66

related interferences to cause complications. Different antibodies weretested without changing the assay epitope specificity as evidenced bypeptidemapping.Clearcross‐reactivityissueswereobservedwithanassayconfigurationutilizingMab‐8I7andMab‐9707together,butnotwhenonlyoneor theotherwasused.Sincecurrenthighsensitivityassaysoftenrelyon different signal amplification methods, which potentially expose theassaystolowaffinity‐relatedinterferences,theissueofantibodyspecificityshould not be ignored even when antibodies with no apparent cross‐reactivitiesarebeingused.

In conclusion, Eu nanoparticles can be employed as the detection system inimmunoassaysrequiringspecificproperties:sensitivityorawidedynamicrangearound a pre‐determined cut‐off value. All the developed assays wereimplementedas5‐or15‐minuteassays,enablingtheirpossibleincorporationintoa POCT assay system. Using recombinant antibody fragments clearly showedsuperiorityoverintactantibodiesbybeingsignificantlylesspronetolow‐affinitymatrix‐related interferences. Additionally, the selection of assay antibodies ishighlightedinthehopeofdevelopinghighlyspecificassaysforcTnI.Thisthesis,therefore,emphasizestheneedtothoroughlystudytheeffectsthattheutilizationofrecombinantantibodyfragmentsandtheselectionofassayantibodieshaveonthe next generation of assays for cTnI. This is supported by the fact that highsensitivity assays have strongly been advocated to beused in the acute cardiactriageandtheevaluationoflongtermcardiacrisk.

Acknowledgements

67

ACKNOWLEDGEMENTS

ThisstudywascarriedoutattheDepartmentofBiotechnologyattheUniversityofTurkuduringtheyears2009–2015.

I am honored to have had the privilege to work with people who have truepassion in what they do. These include Professor Emeritus Timo Lövgren,Professor Kim Pettersson, Professor Tero Soukka and Adjunct Professor UrpoLamminmäki.

Aboveall,IamgratefultomysupervisorsProfessorKimPetterssonandDr.Eeva‐Christine Brockmann. Their support has greatly influenced my growth as ascientist and both have offered importantwords of encouragementwhenmostneeded. I would also like to thank Professor Pettersson for trusting projectmanagement related duties with me: they have enormously developed myprofessionalskills.

I warmly thank allmy co‐authors: Päivi Hedberg, Taina Heikkilä, Henna Kekki,Marja‐Leena Järvenpää,PäiviLaitinen,TarjaPuolakanahoandNooraRistiniemi.In particular, I would like to thank Taina for her diligent work in doing themajorityofpipettingsfortheoriginalpublicationsIIIandIVandforgivingmetheprivilege to guide her. I also wish to thank all the other past workers of theNanoIL‐project I have had the opportunity ofworkingwith: TuomasHuovinen,Tiina Jaatinen, PäiviMalmi and IlariNiemi.Out of all the students and summerworkers in the NanoIL‐project, Laura Mehtälä and Eeva Malmi are especiallythankedfortheircontributionstothis thesiswork.Dr.NooraRistiniemiandDr.Eeva‐ChristineBrockmannarewarmlythankedforreadingthroughthefirstdraftofthework–yourtipsandsuggestionscouldnotbemoreappreciated!

Dr. Susann Eriksson (DHR Finland Oy Innotrac Diagnostics) and Dr. PetriSaviranta (VTT Technical Research Centre of Finland) are courtly thanked forreviewing the thesis work and for giving constructive and valuable comments.ThethanksareextendedtoAnuToivonenforreviewingthelanguageofthethesis.

I wish to thank all the past and present personnel of the Department ofBiotechnology.MirjaJaala,SannaLaitinen,MarjaMaulaandMarttiSointusaloarethankedforkeepingthelaboratoryanditsequipmentupandrunningaswellasfor helpingwithpaperwork. Pirjo Pietilä is thanked for producing the antibodyfragments and Pirjo Laaksonen for sharing an office for a while, as well as forcollectingnumerousbloodsamplesthroughouttheyears.Iwishtoacknowledgeallthepastandpresentdoctoralstudentswhohavesharedtheupsanddownsofacademicresearch,memorablecongresstripsandofferedpricelesspeer‐support.Out of them, Dr. Riina‐Minna Väänänen, Dr. Tanja Savukoski and Dr. HennaPäkkiläaremostwarmlythankedfortheirtips,instructions,anddosanddonotsofthefinalstagesofthePhD‐project.

Acknowledgements

68

Igenuinelythankallmyfriendsandformerteammatesforneverhavingtofullyexplainwhat it is I exactly do atwork. Leaving ourwork lives outside the get‐togethershasprovidedwell‐neededbreaksfromwork.

Mymost sincere andwarmest thanks belong tomy family; mom, dad, andmybrother Kalle and his family. Thank you for always believing in me and myabilitiestogetthedissertationprocessdone–evenwhenIhadmydoubts.

Raisio,February2015

HeidiHyytiä

References

69

REFERENCES

AakreKM,RøraasT,PetersenPH,SvarstadE,SellevollH,SkadbergØ,SæleKandSandbergS.(2014)Weekly and 90‐minute biological variations in cardiac troponin T and cardiac troponin I inhemodialysispatientsandhealthycontrols.ClinChem60:838–847.

AdamSS,KeyNSandGreenbergCS.(2009)D‐dimerantigen:Currentconceptsandfutureprospects.Blood113:2878–2887.

Adams JE, Bodor GS, Dávila‐Román VG, Delmez JA, Apple FS, Ladenson JH and Jaffe AS. (1993)CardiactroponinI.Amarkerwithhighspecificityforcardiacinjury.Circulation88:101–106.

AfifW,LoftusEV,FaubionWA,KaneSV,BruiningDH,HansonKAandSandbornWJ.(2010)Clinicalutilityofmeasuring infliximabandhumananti‐chimericantibodyconcentrations inpatientswithinflammatoryboweldisease.AmJGastroenterol105:1133–1139.

Agewall S, Giannitsis E, Jernberg T and Katus H. (2011) Troponin elevation in coronary vs. non‐coronarydisease.EurHeartJ32:404–411.

Åkerström B and Björck L. (1989) Protein L: An immunoglobulin light chain‐binding bacterialprotein.characterizationofbindingandphysicochemicalproperties.JBiolChem264:19740–19746.

Alpert JS, Thygesen K, Antman E and Bassand JP. (2000) Myocardial infarction redefined‒aconsensus document of the joint european society of cardiology/american college ofcardiologycommitteefortheredefinitionofmyocardialinfarction.JAmCollCardiol36:959–969.

Altshuler EP, Vylegzhanina AV, Katrukha IA, Bereznikova AV and Serebryanaya DV. (2012)Application of recombinant antibody fragments for troponin I measurements.Biochemistry(Mosc)77:1362–1367.

Alvarenga LM, Muzard J, Ledreux A, Bernard C and Billiald P. (2014) Colorimetric engineeredimmunoprobe for the detection and quantification of microcystins. J Immunol Methods406:124–130.

AmundsonBEandAppleFS.(2014)Cardiactroponinassays:Areviewofquantitativepoint‐of‐caredevicesandtheirefficacyinthediagnosisofmyocardialinfarction.ClinChemLabMedOct15,doi:10.1515/cclm‐2014‐0837.[Epubaheadofprint].

AndersenDC,KochC,JensenCH,SkjodtK,BrandtJandTeisnerB.(2004)Highprevalenceofhumananti‐bovine IgG antibodies as the major cause of false positive reactions in two‐siteimmunoassaysbasedonmonoclonalantibodies.JImmunoassayImmunochem25:17–30.

Apple FS, Chung AY, KogutME, Bubany S andMurakamiMM. (2006) Decreased patient chargesfollowing implementation of point‐of‐care cardiac troponin monitoring in acute coronarysyndrome patients in a community hospital cardiology unit.ClinicaChimicaActa370:191–195.

AppleFS,CollinsonPOandIFCCTaskForceonClinicalApplicationsofCardiacBiomarkers.(2012a)Analyticalcharacteristicsofhigh‐sensitivitycardiactroponinassays.ClinChem58:54–61.

AppleFS,LerRandMurakamiMM.(2012b)Determinationof19cardiactroponinIandTassay99thpercentilevaluesfromacommonpresumablyhealthypopulation.ClinChem58:1574–1581.

Armbruster DA and Alexander DB. (2006) Sample to sample carryover: A source of analyticallaboratoryerroranditsrelevancetointegratedclinicalchemistry/immunoassaysystems.ClinChimActa373:37–43.

Baert F, Noman M, Vermeire S, Van Assche G, D'Haens G, Carbonez A and Rutgeerts P. (2003)Influenceofimmunogenicityonthelong‐termefficacyofinfliximabincrohn'sdisease.NEnglJMed348:601–608.

Bale MD, Danielson SJ, Daiss JL, Goppert KE and Sutton RC. (1989) Influence of copolymercomposition on protein adsorption and structural rearrangements at the polymer surface. JColloidandInterfaceSci132:176–187.

References

70

Bale Oenick MD, Danielson SJ, Daiss JL, Sundberg MW and Sutton RC. (1990) Antigen‐bindingactivityofantibodiesimmobilizedonstyrenecopolymerbeads.AnnBiolClin48:651–654.

BallieuxBE,WeijlNI,GelderblomH, vanPelt J andOsanto S. (2008)False‐positive serumhumanchorionicgonadotropin(HCG)inamalepatientwithamalignantgermcelltumorofthetestis:Acasereportandreviewoftheliterature.Oncologist13:1149–1154.

BatesKJ,HallEM,Fahie‐WilsonMN,KindlerH,BaileyC,LythallDandLambEJ.(2010)Circulatingimmunoreactive cardiac troponin forms determined by gel filtration chromatography afteracutemyocardialinfarction.ClinChem56:952–958.

BaumRP,NiesenA,HertelA,NancyA,HessH,DonnerstagB, SykesTR, SykesCJ, SureshMRandNoujaim AA. (1994) Activating anti‐idiotypic human anti‐mouse antibodies forimmunotherapyofovariancarcinoma.Cancer73:1121–1125.

BerglundLandHolmbergNG.(1989)Heterophilicantibodiesagainstrabbitserumcausingfalselyelevatedgonadotropinlevels.ActaObstetGynecolScand68:377‐378.

BergmannO,BhardwajRD,BernardS,ZdunekS,Barnabe‐HeiderF,WalshS,Zupicich J,AlkassK,BuchholzBA,DruidH,etal. (2009)Evidence forcardiomyocyte renewal inhumans.Science324:98–102.

BersonSA,YalowRS,BaumanA,RothschildMAandNewrlyK.(1956)Insulin‐I131metabolisminhumansubjects:Demonstrationofinsulinbindingglobulininthecirculationofinsulintreatedsubjects.JClinInvest35:170–190.

Berth M, Bosmans E, Everaert J, Dierick J, Schiettecatte J, Anckaert E and Delanghe J. (2006)Rheumatoidfactorinterferenceinthedeterminationofcarbohydrateantigen19‐9(CA19‐9).ClinChemLabMed44:1137–1139.

Berthod C and Rey F. (1988) Enormous cross‐reactivity of hydrocortisone hemisuccinate in the"RIANEN"RIAkitforcortisoldetermination.ClinChem34:1358.

BertholfRL,JohannsenLandGuyB.(2002)FalseelevationofserumCA‐125levelcausedbyhumananti‐mouseantibodies.AnnClinLabSci32:414–418.

BetterM, Chang CP, RobinsonRR andHorwitz AH. (1988) Escherichia coli secretion of an activechimericantibodyfragment.Science240:1041–1043.

BingisserR,CairnsC,ChristM,HausfaterP,LindahlB,Mair J,PanteghiniM,PriceCandVengeP.(2012)Cardiactroponin:Acriticalreviewofthecaseforpoint‐of‐caretestingintheED.AmJEmergMed30:1639–1649.

BiondaC,RoussonR,Collin‐ChavagnacD,ManchonM,ChikhKandCharrieA.(2007)UnnecessarycoronaryangiographyduetofalsepositivetroponinIresultsina51‐year‐oldman.ClinChimActa378:225–226.

BjernerJ,NustadK,NorumLF,OlsenKHandBørmerOP.(2002)Immunometricassayinterference:Incidenceandprevention.ClinChem48:613–621.

Bjerner J, Olsen KH, Børmer OP and Nustad K. (2005) Human heterophilic antibodies displayspecificityformurineIgGsubclasses.ClinBiochem38:465–472.

Bjerner J, Bolstad N and Piehler A. (2012) Belief is only half the truth ‐ or why screening forheterophilic antibody interference in certain assaysmakes double sense.Ann ClinBiochem49:381–386.

BlandJMandAltmanDG.(1986)Statisticalmethodsforassessingagreementbetweentwomethodsofclinicalmeasurement.Lancet1:307–310.

BlickK. (2014)Thebenefits of a rapid, point‐of‐care "TnI‐only" zero and2‐hourprotocol for theevaluationofchestpainpatientsintheemergencydepartment.ClinLabMed34:75–85.

BoermanOC,SegersMF,PoelsLG,KenemansPandThomasCM.(1990)Heterophilicantibodiesinhuman sera causing falsely increased results in the CA125 immunofluorometric assay.ClinChem36:888–891.

Bohner J,vonPapeK,HannesWandStegmannT. (1996)False‐negative immunoassayresults forcardiactroponinIprobablyduetocirculatingtroponinIautoantibodies.ClinChem42:2046.

References

71

BolstadN,WarrenDJ,BjernerJ,KravdalG,SchwettmannL,OlsenKH,RustadPandNustadK.(2011)Heterophilic antibody interference in commercial immunoassays; a screening study usingpairednativeandpre‐blockedsera.ClinChemLabMed49:2001–2006.

BolstadN,WarrenDJ andNustad K. (2013)Heterophilic antibody interference in immunometricassays.BestPractResClinEndocrinolMetab27:647–661.

BörmerOP. (1989) Interference of complementwith the binding of carcinoembryonic antigen tosolid‐phasemonoclonalantibodies.JImmunolMethods121:85–93.

Boscato LM and Stuart MC. (1986) Incidence and specificity of interference in two‐siteimmunoassays.ClinChem32:1491–1495.

BoscatoLMandStuartMC. (1988)Heterophilic antibodies:Aproblem for all immunoassays.ClinChem34:27–33.

Boss MA, Kenten JH, Wood CR and Emtage JS. (1984) Assembly of functional antibodies fromimmunoglobulinheavyandlightchainssynthesisedinE.coli.NucleicAcidsRes12:3791–3806.

BounameauxH,PerrierAandRighiniM.(2010)Diagnosisofvenousthromboembolism:Anupdate.VascMed15:399–406.

Brenner B, Francis CW, Totterman S, Kessler CM, Rao AK, Rubin R, Kwaan HC, Gabriel KR andMarder VJ. (1990) Quantitation of venous clot lysis with the D‐dimer immunoassay duringfibrinolytic therapy requires correction for soluble fibrin degradation.Circulation81:1818–1825.

BrinkleyM.(1992)Abriefsurveyofmethodsforpreparingproteinconjugateswithdyes,haptens,andcross‐linkingreagents.BioconjugChem3:2–13.

BrockmannEC,VehniäinenMandPetterssonK.(2010)Useofhigh‐capacitysurfacewithorientedrecombinant antibody fragments in a 5‐min immunoassay for thyroid‐stimulating hormone.AnalBiochem396:242–249.

BrockmannEC,AkterS,SavukoskiT,HuovinenT,LehmusvuoriA,LeivoJ,SaavalainenO,AzhayevA,Lövgren T, Hellman J, et al. (2011) Synthetic single‐framework antibody library integratedwithrapidaffinitymaturationbyVLshuffling.ProteinEngDesSel24:691–700.

Butch AW. (2000) Falsely increased values for rabbit immunoglobulin‐based nephelometricimmunoassaysattributabletohumananti‐rabbitantibodies.ClinChem46:1207–1208.

ButlerSAandColeLA.(2001)Useofheterophilicantibodyblockingagent(HBT)inreducingfalse‐positivehCGresults.ClinChem47:1332–1333.

CampbellRD,DoddsAWandPorterRR.(1980)ThebindingofhumancomplementcomponentC4toantibody‐antigenaggregates.BiochemJ189:67–80.

CarlanderDandLarssonA.(2001)AvianantibodiescaneliminateinterferenceduetocomplementactivationinELISA.UpsJMedSci106:189–195.

CarsonDA,ChenPP,FoxRI,KippsTJ,JirikF,GoldfienRD,SilvermanG,RadouxVandFongS.(1987)Rheumatoidfactorandimmunenetworks.AnnuRevImmunol5:109–126.

CasalsG,FilellaXandBediniJL.(2007)EvaluationofanewultrasensitiveassayforcardiactroponinI.ClinBiochem40:1406–1413.

Cavalier E, Delanaye P, Carlisi A, Chapelle J and Collette J. (2009) An unusual interference inparathormoneassaycausedbyanti‐goatIgG:Acasereport.ClinChemLabMed47:118.

ChangCY,Lu JY,ChienTI,Kao JT,LinMC,ShihPCandYanSN. (2003) InterferencecausedbythecontentsofserumseparatortubesinthevitrosCRPassay.AnnClinBiochem40:249–251.

ChapmanDW,SimpsonJA,IscoeS,RobinsTandNosakaK.(2013)Changesinserumfastandslowskeletal troponin I concentration followingmaximal eccentric contractions. J SciMed Sport16:82–85.

Check JH, Ubelacker L and Lauer CC. (1995) Falsely elevated steroidal assay levels related toheterophileantibodiesagainstvariousanimalspecies.GynecolObstetInvest40:139–140.

Chen M, Serizawa T, Kishida A and Akashi M. (1999) Graft copolymers having hydrophobicbackboneandhydrophilicbranches.XXIII.particlesizecontrolofpoly(ethyleneglycol)‐coated

References

72

polystyrene nanoparticles prepared bymacromonomermethod. JAppl Polym Sci37:2155–2166.

ChitarraV,AlzariPM,BentleyGA,BhatTN,EiseleJL,HoudusseA,LescarJ,SouchonHandPoljakRJ.(1993)Three‐dimensionalstructureofaheterocliticantigen‐antibodycross‐reactioncomplex.ProcNatlAcadSciUSA90:7711–7715.

ChristM,BertschT,PoppS,BahrmannP,HeppnerHJandMüllerC.(2011)High‐sensitivitytroponinassays intheevaluationofpatientswithacutechestpain intheemergencydepartment.ClinChemLabMed49:1955–1963.

ChristensonRH,DuhSH,AppleFS,BodorGS,BunkDM,PanteghiniM,WelchMJ,WuAHandKahnSE. (2006) Toward standardization of cardiac troponin I measurements part II: Assessingcommutability of candidate reference materials and harmonization of cardiac troponin Iassays.ClinChem52:1685–1692.

Clinical and Laboratory Standards Institute. (2004) CLSI; evaluation of precision performance ofquantitativemeasurementmethods;approvedguideline‐secondedition.CLSIdocumentEP5‐A2.Wayne,PA.

Clinical and Laboratory Standards Institute. (2012) CLSI; evaluation of detection capability forclinical laboratory measurement procedures; approved guideline‐second edition. CLSIdocumentEP17‐A2.Wayne,PA.

ColeLA,ShahabiS,ButlerSA,MitchellH,NewlandsES,BehrmanHRandVerrillHL.(2001)Utilityofcommonly used commercial human chorionic gonadotropin immunoassays in the diagnosisandmanagementoftrophoblasticdiseases.ClinChem47:308–315.

ColnotDR,QuakJJ,RoosJC,vanLingenA,WilhelmAJ,vanKampGJ,HuijgensPC,SnowGBandvanDongenGA.(2000)PhaseItherapystudyof186Re‐labeledchimericmonoclonalantibodyU36inpatientswithsquamouscellcarcinomaoftheheadandneck.JNuclMed41:1999–2010.

CovinskyM,LaterzaO,PfeiferJD,Farkas‐SzallasiTandScottMG.(2000)AnIgMlambdaantibodytoescherichia coli produces false‐positive results inmultiple immunometric assays.ClinChem46:1157–1161.

Csako G,Weintraub BD and Zweig MH. (1988) The potency of immunoglobulin G fragments forinhibition of interference caused by anti‐immunoglobulin antibodies in a monoclonalimmunoradiometricassayforthyrotropin.ClinChem34:1481–1483.

CumminsB,AucklandMLandCumminsP.(1987)Cardiac‐specifictroponin‐Iradioimmunoassayinthediagnosisofacutemyocardialinfarction.AmHeartJ113:1333–1344.

Dahlmann N and Bidlingmaier F. (1989) Circulating antibodies to mouse monoclonalimmunoglobulins caused false‐positive results in a two‐site assay for alpha‐fetoprotein.ClinChem35:2339.

Dasgupta A, Banerjee SK and Datta P. (1999) False‐positive troponin I in the MEIA due to thepresenceofrheumatoidfactorsinserum.Eliminationofthisinterferencebyusingapolyclonalantiseraagainstrheumatoidfactors.AmJClinPathol112:753–756.

DasguptaA,BiddleDA,WellsAandDattaP.(2000)Positiveandnegativeinterferenceofthechinesemedicine chan su in serum digoxin measurement. Elimination of interference by using amonoclonal chemiluminescentdigoxin assayormonitoring freedigoxin concentration.Am JClinPathol114:174–179.

DasguptaA,WellsAandBiddleD.(2001)NegativeinterferenceofbilirubinandhemoglobinintheMEIAtroponinIassaybutnotintheMEIACK‐MBassay.JClinLabAnal15:76–80.

DasguptaA.(2006)Therapeuticdrugmonitoringofdigoxin:Impactofendogenousandexogenousdigoxin‐likeimmunoreactivesubstances.ToxicolRev25:273–281.

Dasgupta A and Bernard DW. (2006) Herbal remedies: Effects on clinical laboratory tests. ArchPatholLabMed130:521–528.

Davidsohn I and Lee CL. (1966) Heterophilic antibodies following injection of blood groupsubstances.Transfusion6:487–492.

Davies C. (2013) Immunoassay performance measures. In: Wild D. (edited), The ImmunoassayHandbook112–26.ElsevierScience,TheNetherlands.

References

73

de Jager W, Prakken BJ, Bijlsma JW, Kuis W and Rijkers GT. (2005) Improved multipleximmunoassay performance in human plasma and synovial fluid following removal ofinterferingheterophilicantibodies.JImmunolMethods300:124–135.

DeForge LE, Loyet KM, Delarosa D, Chinn J, Zamanian F, Chuntharapai A, Lee J, Hass P, Wei N,TownsendMJ, etal. (2010)Evaluation of heterophilic antibody blocking agents in reducingfalse positive interference in immunoassays for IL‐17AA, IL‐17FF, and IL‐17AF. J ImmunolMethods362:70–81.

Demers L and SpencerC. (2003) Laboratory support for thediagnosis andmonitoring of thyroiddisease:Thyroglobulin(tg)measurement.Thyroid13:57–67.

DespresN andGrant AM. (1998)Antibody interference in thyroid assays: A potential for clinicalmisinformation.ClinChem44:440–454.

DezelićG,DezelićNandTelismanZ.(1971)Thebindingofhumanserumalbuminbymonodispersepolystyrenelatexparticles.EurJBiochem23:575–581.

DiamandisEP,MortonRC,ReichsteinEandKhosraviMJ.(1989)Multiplefluorescencelabelingwitheuropiumchelators.applicationtotime‐resolvedfluoroimmunoassays.AnalChem61:48–53.

DiamandisEPandChristopoulosTK.(1990)Europiumchelatelabelsintime‐resolvedfluorescenceimmunoassaysandDNAhybridizationassays.AnalChem62:1149A–1157A.

Dorizzi RM, Tagliaro F, Ghielmi S, Archetti S, Poiesi C and Luisetto G. (1991) Antibodies againstcalcitoninsasasourceofanalyticalerrors.EurJClinChemClinBiochem29:689.

Drain PK, Hyle EP, Noubary F, Freedberg KA,WilsonD, BishaiWR, RodriguezW and Bassett IV.(2014)Diagnostic point‐of‐care tests in resource‐limited settings.Lancet InfectDis14:239–249.

DrenckN.(2001)Pointofcaretestingincriticalcaremedicine:Theclinician'sview.ClinChimActa307:3–7.

Eggers KM, Venge P and Lindahl B. (2012) High‐sensitive cardiac troponin T outperforms noveldiagnosticbiomarkersinpatientswithacutechestpain.ClinChimActa413:1135–1140.

EggersKM,VengeP,LindahlBandLindL.(2013)CardiactroponinI levelsmeasuredwithahigh‐sensitive assay increase over time and are strong predictors of mortality in an elderlypopulation.JAmCollCardiol61:1906–1913.

EichingerS.(2005)D‐dimertestinginpregnancy.SeminVascMed5:375–378.EkinsRP.(1960)Theestimationofthyroxineinhumanplasmabyanelectrophoretictechnique.Clin

ChimActa5:453–459.EmersonJF,NgoGandEmersonSS.(2003)Screeningforinterferenceinimmunoassays.ClinChem

49:1163–1169.Eriksson S, VehniäinenM, JansenT,MeretojaV, Saviranta P, PetterssonK and LövgrenT. (2000)

Dual‐labeltime‐resolvedimmunofluorometricassayoffreeandtotalprostate‐specificantigenbasedonrecombinantfabfragments.ClinChem46:658–666.

ErikssonS, JunikkaM, LaitinenP,Majamaa‐VolttiK,AlfthanH andPetterssonK. (2003)Negativeinterference in cardiac troponin I immunoassays from a frequently occurring serum andplasmacomponent.ClinChem49:1095–1104.

Eriksson S, Halenius H, Pulkki K, Hellman J and Pettersson K. (2005a) Negative interference incardiac troponin I immunoassaysbycirculating troponinautoantibodies.ClinChem51:839–847.

ErikssonS,IlvaT,BeckerC,LundJ,PorelaP,PulkkiK,Voipio‐PulkkiLMandPetterssonK.(2005b)Comparison of cardiac troponin I immunoassays variably affected by circulatingautoantibodies.ClinChem51:848–855.

Evans MJ, Livesey JH, Ellis MJ and Yandle TG. (2001) Effect of anticoagulants and storagetemperaturesonstabilityofplasmaandserumhormones.ClinBiochem34:107–112.

Fahie‐WilsonMandSmithTP.(2013)Determinationofprolactin:Themacroprolactinproblem.BestPractResClinEndocrinolMetab27:725–742.

References

74

FalchukKRandIsselbacherKJ.(1976)Circulatingantibodiestobovinealbumininulcerativecolitisandcrohn'sdisease.characterizationoftheantibodyresponse.Gastroenterology70:5–8.

FarahCSandReinachFC.(1995)Thetroponincomplexandregulationofmusclecontraction.FASEBJ9:755–767.

FermannGJandSuyamaJ.(2002)Pointofcaretestingintheemergencydepartment.JEmergMed22:393–404.

FitzmauriceTF,BrownC,RifaiN,WuAHandYeoKT. (1998)False increaseof cardiac troponin Iwithheterophilicantibodies.ClinChem44:2212–2214.

FlemingSM,O'ByrneL,FinnJ,GrimesHandDalyKM.(2002)False‐positivecardiactroponinIinaroutineclinicalpopulation.AmJCardiol89:1212–1215.

FrengenJ,KierulfB,SchmidR,LindmoTandNustadK.(1994)Demonstrationandminimizationofseruminterferenceinflowcytometrictwo‐siteimmunoassays.ClinChem40:420–425.

Fritz BE, Hauke RJ and Stickle DF. (2009) New onset of heterophilic antibody interference inprostate‐specific antigen measurement occurring during the period of post‐prostatectomyprostate‐specificantigenmonitoring.AnnClinBiochem46:253–256.

FrödinJE,LefvertAKandMellstedtH.(1992)TheclinicalsignificanceofHAMAinpatientstreatedwithmousemonoclonalantibodies.CellBiophys21:153–165.

GallagherDJ,RichesJandBajorinDF.(2010)Falseelevationofhumanchorionicgonadotropininapatientwithtesticularcancer.NatRevUrol7:230–233.

García‐Mancebo ML, Agulló‐Ortuño MT, Gimeno JR, Navarro‐Martínez MD, Ruíz‐Gómez J andNoguera‐Velasco JA. (2005) Heterophile antibodies produce spuriously elevatedconcentrations of cardiac troponin I in patients with legionella pneumophila. Clin Biochem38:584–587.

Gaze DC and Collinson PO. (2008) Multiple molecular forms of circulating cardiac troponin:Analyticalandclinicalsignificance.AnnClinBiochem45:349–355.

Ghali S,LewisK,KazanV,AltorokN,Taji J,TalebM,LankaKandAssalyR. (2012)FluctuationofspuriouslyelevatedtroponinI:Acasereport.CaseRepCritCare2012:585879.

Giannitsis E, BeckerM,KurzK,HessG, ZdunekD andKatusHA. (2010a)High‐sensitivity cardiactroponinTforearlypredictionofevolvingnon‐ST‐segmentelevationmyocardialinfarctioninpatientswithsuspectedacutecoronarysyndromeandnegativetroponinresultsonadmission.ClinChem56:642–650.

GiannitsisE,KurzK,HallermayerK,JarauschJ,JaffeASandKatusHA.(2010b)Analyticalvalidationofahigh‐sensitivitycardiactroponinTassay.ClinChem56:254–261.

Giuliani I, Bertinchant JP, Granier C, Laprade M, Chocron S, Toubin G, Etievent JP, Larue C andTrinquier S. (1999)Determinationof cardiac troponin I forms in thebloodofpatientswithacutemyocardialinfarctionandpatientsreceivingcrystalloidorcoldbloodcardioplegia.ClinChem45:213–222.

Gonzales N. (2003)Minimizing immunogenicity of the SDR‐grafted humanized antibody CC49 bygeneticmanipulationoftheframeworkresidues.MolImmunol40:337–349.

GoodacreS,CrossE,ArnoldJ,AngeliniK,CapewellSandNichollJ.(2005)Thehealthcareburdenofacutechestpain.Heart91:229–230.

GoodmanGE,BeaumierP,HellströmI,FernyhoughBandHellströmKE.(1985)Pilottrialofmurinemonoclonalantibodiesinpatientswithadvancedmelanoma.JClinOncol3:340–352.

Griffin C, Sutor J and Shull B. (1994) Introduction. In Microparticle Reagent Optimization: ALaboratoryReferenceManualfromtheAuthorityonMicroparticles11–26.Seradyn,IN,UnitedStates.

GrilloR,RosaAandFracetoL.(2014)Engineerednanoparticlesandorganicmatter:Areviewofthestate‐of‐the‐art.Chemosphere119C:608–619.

Hagan AK and Zuchner T. (2011) Lanthanide‐based time‐resolved luminescence immunoassays.AnalBioanalChem400:2847–2864.

References

75

Hakala H, Mukkala V, Sutela T and Hovinen J. (2006) Synthesis and properties of nanospherescopolymerisedwithluminescenteuropium(iii)chelates.OrgBiomolChem4:1383–1386.

HammCW,GiannitsisEandKatusHA.(2002)Cardiactroponinelevationsinpatientswithoutacutecoronarysyndrome.Circulation106:2871–2872.

Harkiss GD. (1984) Antibodies to equine antithymocyte globulin in heart transplant recipients:Evaluationofanenzymeimmunoassay.JClinLabImmunol15:175–180.

HärmäH,SoukkaTandLövgrenT.(2001)Europiumnanoparticlesandtime‐resolvedfluorescenceforultrasensitivedetectionofprostate‐specificantigen.ClinChem47:561–568.

HarveyMA.(1999)Point‐of‐carelaboratorytestingincriticalcare.AmJCritCare8:72–83.HastingsKE.(1997)MolecularevolutionofthevertebratetroponinIgenefamily.CellStructFunct

22:205–211.HawkinsRC. (2003)Hemolysis interference in theortho‐clinicaldiagnosticsvitrosECicTnIassay.

ClinChem49:1226–1227.HaydenMS,GillilandLKandLedbetterJA.(1997)Antibodyengineering.CurrOpinImmunol9:201–

212.HeitJA,SilversteinMD,MohrDN,PettersonTM,LohseCM,O'FallonWMandMeltonLJ,3rd.(2001)

Theepidemiologyofvenousthromboembolisminthecommunity.ThrombHaemost86:452–463.

Hennig C, Rink L, Fagin U, JabsWJ and Kirchner H. (2000) The influence of naturally occurringheterophilicanti‐immunoglobulinantibodiesondirectmeasurementofserumproteinsusingsandwichELISAs.JImmunolMethods235:71–80.

HickmanPE,Potter JM,AroneyC,KoerbinG,SouthcottE,WuAHandRobertsMS. (2010)Cardiactroponinmaybereleasedbyischemiaalone,withoutnecrosis.ClinChimActa411:318–323.

HiemstraPS,BaldwinWM,vanderVoortEA,PaulLC,vanEsLAandDahaMR.(1988)PolymericIgAantibody response to rabbit antithymocyte globulin in renal transplant recipients.Transplantation45:701–705.

HolmesKLandLantzLM.(2001)Proteinlabelingwithfluorescentprobes.MethodsCellBiol63:185–204.

Hu X, O'Hara L,White S,Magner E, KaneM andWall JG. (2007) Optimisation of production of adomoic acid‐binding scFvantibody fragment in escherichia coliusingmolecular chaperonesand functional immobilisation on amesoporous silicate support.ProteinExprPurif52:194‐201.

Huckle D. (2013) Market trends. In: Wild D. (edited), The Immunoassay Handbook 517–531.ElsevierScience,TheNetherlands.

HudsonPJ.(1998)Recombinantantibodyfragments.CurrOpinBiotechnol9:395–402.Hueber W, Tomooka BH, Zhao X, Kidd BA, Drijfhout JW, Fries JF, van Venrooij WJ, Metzger AL,

Genovese MC and Robinson WH. (2007) Proteomic analysis of secreted proteins in earlyrheumatoid arthritis: Anti‐citrulline autoreactivity is associated with up regulation ofproinflammatorycytokines.AnnRheumDis66:712–719.

HuhtinenP,VaarnoJ,SoukkaT,LövgrenTandHärmäH.(2004)Europium(III)nanoparticle‐label‐basedassayforthedetectionofnucleicacids.Nanotechnology15:1708–1715.

HuhtinenP,KiveläM,KuronenO,HagrenV,TakaloH,TenhuH, LövgrenT andHärmäH. (2005)Synthesis, characterization, and application of eu(III), tb(III), sm(III), and dy(III) lanthanidechelatenanoparticlelabels.AnalChem77:2643–2648.

HunterWM and Budd PS. (1980) Circulating antibodies to ovine and bovine immunoglobulin inhealthysubjects:Ahazardforimmunoassays.Lancet2:1136.

HwangWYandFooteJ.(2005)Immunogenicityofengineeredantibodies.Methods36:3‐10.IsmaN, Svensson PJ, Gottsäter A and Lindblad B. (2009) Prospective analysis of risk factors and

distributionofvenousthromboembolisminthepopulation‐basedmalmöthrombophiliastudy(MATS).ThrombRes124:663–666.

References

76

Ismail AA. (2007) On detecting interference from endogenous antibodies in immunoassays bydoublingdilutionstest.ClinChemLabMed45:851–854.

JaffeAS,VasileVC,MiloneM,SaengerAK,OlsonKNandAppleFS.(2011)Diseasedskeletalmuscle:Anoncardiac source of increased circulating concentrations of cardiac troponin T. J Am CollCardiol58:1819–1824.

JaneEllisM,LiveseyJHandEvansMJ.(2003)Hormonestabilityinhumanwholeblood.ClinBiochem36:109–112.

JanssenMJW,VelmansMHandHeesenWF. (2014)Apatientwith ahigh concentrationofB‐typenatriuretic peptide (BNP) but normalN‐terminal proBNP concentration: A case report.ClinBiochem47:1136–1137.

JärvenpääML,KuningasK,Niemi I,HedbergP,RistiniemiN, PetterssonK andLövgrenT. (2012)Rapid and sensitive cardiac troponin I immunoassay based on fluorescent europium(III)‐chelate‐dyednanoparticles.ClinChimActa414:70–75.

JespersLS,RobertsA,MahlerSM,WinterGandHoogenboomHR. (1994)Guiding theselectionofhuman antibodies from phage display repertoires to a single epitope of an antigen.Biotechnology(NY)12:899–903.

Ji JZ andMeng QH. (2011) Evaluation of the interference of hemoglobin, bilirubin, and lipids onrochecobas6000assays.ClinChimActa412:1550–1553.

KaplanIVandLevinsonSS.(1999)Whenisaheterophileantibodynotaheterophileantibody?whenitisanantibodyagainstaspecificimmunogen.ClinChem45:616–618.

Kato K, Umeda U, Suzuki F, Hayashi D and Kosaka A. (1979) Use of antibody fab' fragments toremove interferencebyrheumatoid factorswith theenzyme‐linkedsandwich immunoassay.FEBSLett102:253–256.

KatrukhaAG,BereznikovaAV,EsakovaTV,PetterssonK,LövgrenT,SeverinaME,PulkkiK,Vuopio‐PulkkiLMandGusevNB.(1997)TroponinIisreleasedinbloodstreamofpatientswithacutemyocardialinfarctionnotinfreeformbutascomplex.ClinChem43:1379–1385.

KatusHA,LooserS,HallermayerK,RemppisA,ScheffoldT,BorgyaA,EssigUandGeussU.(1992)DevelopmentandinvitrocharacterizationofanewimmunoassayofcardiactroponinT.ClinChem38:386–393.

Kazmierczak SC, Sekhon H and Richards C. (2005) False‐positive troponin I measured with theabbottAxSYMattributedtofibrininterference.IntJCardiol101:27–31.

KellerT,ZellerT,PeetzD,TzikasS,RothA,CzyzE,BickelC,BaldusS,WarnholtzA,FrohlichM,etal.(2009)Sensitive troponin I assay in earlydiagnosis of acutemyocardial infarction.NEngl JMed361:868–877.

Kelly LS, KozakM,Walker T, PierceM and Puett D. (2005) Lectin immunoassays using antibodyfragments to detect glycoforms of human chorionic gonadotropin secreted bychoriocarcinomacells.AnalBiochem338:253–262.

KennaJG,MajorGNandWilliamsRS.(1985)Methodsforreducingnon‐specificantibodybindinginenzyme‐linkedimmunosorbentassays.JImmunolMethods85:409–419.

KerschbaumerRJ,Hirschl S, KaufmannA, IblM, KoenigR andHimmler G. (1997) Single‐chain fvfusion proteins suitable as coating and detecting reagents in a double antibody sandwichenzyme‐linkedimmunosorbentassay.AnalBiochem249:219–227.

KhaliliHanddeLemosJA. (2014)Whatconstitutesarelevantchange inhigh‐sensitivity troponinvaluesoverserialmeasurement?ClinChem60:803–805.

KielyPDW.(2000)Serumskeletal troponinI in inflammatorymuscledisease:Relationtocreatinekinase,CKMBandcardiactroponinI.AnnRheumDis59:750–751.

KimS,YunYM,HurM,MoonHWandKimJQ.(2011)Theeffectsofanti‐insulinantibodiesandcross‐reactivity with human recombinant insulin analogues in the E170 insulin immunometricassay.KoreanJLabMed31:22–29.

References

77

Kim WJ, Laterza OF, Hock KG, Pierson‐Perry JF, Kaminski DM, Mesguich M, Braconnier F,Zimmermann R, Zaninotto M, Plebani M, et al. (2002) Performance of a revised cardiactroponin method that minimizes interferences from heterophilic antibodies. Clin Chem48:1028–1034.

Knappik A and Plückthun A. (1994) An improved affinity tag based on the FLAG peptide for thedetectionandpurificationofrecombinantantibodyfragments.BioTechniques17:754–761.

Koerbin G, AbhayaratnaWP, Potter JM, Apple FS, Jaffe AS, Ravalico TH and Hickman PE. (2013)Effectofpopulationselectionon99thpercentilevaluesforahighsensitivitycardiactroponinIandTassays.ClinBiochem46:1636–1643.

KöhlerGandMilsteinC.(1975)Continuousculturesoffusedcellssecretingantibodyofpredefinedspecificity.Nature256:495–497.

Korley FK and Jaffe AS. (2013) Preparing the united states for high‐sensitivity cardiac troponinassays.JAmCollCardiol61:1753–1758.

Koshida S, AsanumaK, KuribayashiK, GotoM, TsujiN, KobayashiD, TanakaM andWatanabeN.(2010) Prevalence of human anti‐mouse antibodies (HAMAs) in routine examinations. ClinChimActa411:391–394.

Kricka LJ. (1999) Human anti‐animal antibody interferences in immunological assays. Clin Chem45:942–956.

Krintus M, Kozinski M, Boudry P, Capell NE, Koller U, Lackner K, Lefevre G, Lennartz L, Lotz J,Herranz AM, et al. (2014) European multicenter analytical evaluation of the abbottARCHITECTSTAThighsensitivetroponinIimmunoassay.ClinChemLabMed52:1657–1665.

Kroll MH and Elin RJ. (1994) Interference with clinical laboratory analyses. Clin Chem40:1996–2005.

KurokiM,MatsumotoY,ArakawaF,HarunoM,MurakamiM,KuwaharaM,OzakiH, SenbaT andMatsuoka Y. (1995) Reducing interference from heterophilic antibodies in a two‐siteimmunoassay for carcinoembryonic antigen (CEA) by using a human/mouse chimericantibodytoCEAasthetracer.JImmunolMethods180:81–91.

Kuus‐Reichel K, Grauer LS, Karavodin LM, Knott C, Krusemeier M and Kay NE. (1994) Willimmunogenicity limit the use, efficacy, and future development of therapeutic monoclonalantibodies?ClinDiagnLabImmunol1:365–372.

Lappé JM, Pelfrey CM, Cotleur A and TangWH. (2011) Cellular proliferative response to cardiactroponin‐Iinpatientswithidiopathicdilatedcardiomyopathy.ClinTranslSci4:317–322.

Larsson A, Hedenborg G and Carlström A. (1981) Placental transfer of maternal anti‐rabbit IgGcausingfalselyelevatedTSHlevelsinneonates.ActaPaediatrScand70:699–703.

Larue C, Calzolari C, Bertinchant JP, Leclercq F, Grolleau R and Pau B. (1993) Cardiac‐specificimmunoenzymometric assayof troponin I in theearlyphaseof acutemyocardial infarction.ClinChem39:972–979.

Laukkanen ML, Orellana A and Keinänen K. (1995) Use of genetically engineered lipid‐taggedantibody to generate functional europium chelate‐loaded liposomes. application influoroimmunoassay.JImmunolMethods185:95–102.

LeGattDF,BlakneyGB,HigginsTN,SchnablKL,ShalapayCE,DiasVCandWesenbergJC.(2012)Theeffect of paraproteins and rheumatoid factor on four commercial immunoassays forvancomycin: Implications for laboratorians and other health care professionals. Ther DrugMonit34:306–311.

LevineS,NothR,LooA,ChopraIJandKleeGG.(1990)AnomalousserumthyroxinmeasurementswiththeabbottTDxprocedure.ClinChem36:1838–1840.

Levinson SS and Miller JJ. (2002) Towards a better understanding of heterophile (and the like)antibodyinterferencewithmodernimmunoassays.ClinChimActa325:1–15.

LiY,WangX,XuLandWenX.(2014)RapididentificationoffalselyelevatedserumcardiactroponinIvaluesinastatlaboratory.LabMed45:82–85.

References

78

LiddellE. (2013)Antibodies. In:WildD. (edited),The ImmunoassayHandbook245–265.ElsevierScience,TheNetherlands.

LiendoC,GhaliJKandGravesSW.(1996)Anewinterferenceinsomedigoxinassays:Anti‐murineheterophilicantibodies.ClinPharmacolTher60:593–598.

LindP,LechnerP,Arian‐SchadK,KlimpfingerM,CesnikH,KammerhuberFandEberO.(1991)Anti‐carcinoembryonic antigen immunoscintigraphy (technetium‐99m‐monoclonal antibody BW431/26) and serumCEA levels in patientswith suspectedprimary and recurrent colorectalcarcinoma.JNuclMed32:1319–1325.

LindahlB,VengePandJamesS.(2010)Thenewhigh‐sensitivitycardiactroponinTassayimprovesriskassessmentinacutecoronarysyndromes.AmHeartJ160:224–229.

LindnerP,BauerK,KrebberA,NiebaL,KremmerE,KrebberC,HoneggerA,KlingerB,MocikatRandPlückthunA. (1997) Specific detection of his‐tagged proteinswith recombinant anti‐his tagscFv‐phosphataseorscFv‐phagefusions.BioTechniques22:140–149.

LindsellCJ,AnantharamanV,DiercksD,HanJH,HoekstraJW,HollanderJE,KirkJD,LimSH,PeacockWF, Tiffany B, et al. (2006) The internet tracking registry of acute coronary syndromes(i*trACS): A multicenter registry of patients with suspicion of acute coronary syndromesreported using the standardized reporting guidelines for emergency department chest painstudies.AnnEmergMed48:666–677.

Lipman NS, Jackson LR, Trudel LJ and Weis‐Garcia F. (2005) Monoclonal versus polyclonalantibodies: Distinguishing characteristics, applications, and information resources. ILAR J46:258–268.

Lippi G, Aloe R, Meschi T, Borghi L and Cervellin G. (2013a) Interference from heterophilicantibodies in troponintesting.casereportandsystematicreviewof the literature.ClinChimActa426:79–84.

LippiG,FranchiniMandCervellinG.(2013b)Diagnosisandmanagementofischemicheartdisease.SeminThrombHemost39:202–213.

LippiG,IppolitoL,TondelliMTandFavaloroEJ.(2014)InterferencefromheterophilicantibodiesinD‐dimerassessment.Acasereport.BloodCoagulFibrinolysis25:277–279.

Liu AY, Robinson RR, Hellström KE, Murray ED,Jr, Chang CP and Hellström I. (1987) Chimericmouse‐human IgG1antibody that canmediate lysisof cancer cells.ProcNatlAcadSciUSA84:3439–3443.

LohTP,KaoSL,HalsallDJ,TohSA,ChanE,HoSC,TaiESandKhooCM.(2012)Macro‐thyrotropin:Acasereportandreviewofliterature.JClinEndocrinolMetab97:1823–1828.

Lövestam G, Rauscher H, Roebben G, Klüttgen B, Gibson N, Putaud J and Stamm H. (2010)Considerationsonadefinitionofnanomaterialforregulatorypurposes.PublicationsOfficeoftheEuropeanUnion,Luxembourg.

LövgrenT,HemmiläI,PetterssonK,EskolaJUandBertoftE.(1984)Determinationofhormonesbytime‐resolvedfluoroimmunoassay.Talanta31:909–916.

Lück M, Paulke BR, Schröder W, Blunk T and Müller RH. (1998) Analysis of plasma proteinadsorptiononpolymericnanoparticleswithdifferentsurfacecharacteristics. JBiomedMaterRes39:478–485.

LumG, SolarzD and Farney L. (2006) False positive cardiac troponin results in patientswithoutacutemyocardialinfarction.LabMed37:546–550.

LyonsA,KingDJ,OwensRJ,YarrantonGT,MillicanA,WhittleNRandAdairJR.(1990)Site‐specificattachment to recombinant antibodies via introduced surface cysteine residues.ProteinEng3:703–708.

MarianiM,CamagnaM,TarditiLandSeccamaniE.(1991)Anewenzymaticmethodtoobtainhigh‐yieldF(ab)2suitableforclinicalusefrommouseIgGl.MolImmunol28:69–77.

Marks V. (2002) False‐positive immunoassay results: A multicenter survey of erroneousimmunoassayresults fromassaysof74analytes in10donors from66 laboratories insevencountries.ClinChem48:2008–2016.

References

79

Martel J,DespresN,AhnadiCE,Lachance JF,Monticello JE,FinkG,ArdemagniA,BanfiG,Tovey J,DykesP,etal.(2000)Comparativemulticentrestudyofapanelofthyroidtestsusingdifferentautomatedimmunoassayplatformsandspecimensathighriskofantibodyinterference.ClinChemLabMed38:785–793.

MasimasiNandMeansRTJ.(2005)Elevatedtroponinlevelsassociatedwithhemolysis.AmJMedSci330:201–203.

Matsuya T, Tashiro S, HoshinoN, ShibataN, Nagasaki Y andKataokaK. (2003) A core‐shell‐typefluorescent nanosphere possessing reactive poly(ethylene glycol) tethered chains on thesurface forzeptomoledetectionofprotein in time‐resolved fluorometric immunoassay.AnalChem75:6124–6132.

McClennen S, Halamka J, Horowitz G, Kannam J and Ho K. (2003) Clinical prevalence andramificationsof false‐positivecardiactroponinIelevationsfromtheabbottAxSYManalyzer.AmJCardiol91:1125–1127.

McNeelDG,NguyenLD,StorerBE,VessellaR,LangePHandDisisML.(2000)Antibodyimmunitytoprostate cancer associated antigens can be detected in the serum of patientswith prostatecancer.JUrol164:1825–1829.

Melanson S, Conrad M, Mosammaparast N and Jarolim P. (2008) Implementation of a highlysensitivecardiactroponinIassay:Testvolumes,positivityratesandinterpretationofresults.ClinChimActa395:57–61.

MonchampT,ChopraI,WahDandButchA.(2007)Falselyelevatedthyroidhormonelevelsduetoanti‐sheep antibody interference in an automated electro chemiluminescent immunoassay.Thyroid17:271–275.

MorrisonAE,LudlamCAandKesslerC.(1993)UseofporcinefactorVIIIinthetreatmentofpatientswithacquiredhemophilia.Blood81:1513–1520.

MorrowDAandAntmanEM.(2009)Evaluationofhigh‐sensitivityassaysforcardiactroponin.ClinChem55:5–8.

Moseley KR, Knapp RC and Haisma HJ. (1988) An assay for the detection of human anti‐murineimmunoglobulinsinthepresenceofCA125antigen.JImmunolMethods106:1–6.

Mössner E, Lenz H and Bienhaus G. (1990) Elimination of heterophilic antibody interference inmonoclonalsandwichtests[abstract].ClinChem36:1093.

MuellerC.(2014)Biomarkersandacutecoronarysyndromes:Anupdate.EurHeartJ35:552‐556.Mukkala VM, Mikola H and Hemmilä I. (1989) The synthesis and use of activated N‐benzyl

derivatives of diethylenetriaminetetraacetic acids: Alternative reagents for labeling ofantibodieswithmetalions.AnalBiochem176:319–325.

Müller W, Mierau R and Wohltmann D. (1985) Interference of IgM rheumatoid factor withnephelometricC‐reactiveproteindeterminations.JImmunolMethods80:77–90.

Nichols JH. (2013) Point‐of‐care testing. In: Wild D. (edited), The Immunoassay Handbook, 4thedition455–463.ElsevierScience,TheNetherlands.

NicholsonS,FoxM,EpenetosAandRustinG.(1996)Immunoglobulininhibitingreagent:Evaluationof a newmethod for eliminating spurious elevations in CA125 caused by HAMA. Int JBiolMarkers11:46–49.

Nielsen SF, Bojesen SE, Schnohr P and Nordestgaard BG. (2012) Elevated rheumatoid factor andlongtermriskofrheumatoidarthritis:Aprospectivecohortstudy.BMJ345:e5244.

NilssonUR,StormKE,ElwingHandNilssonB.(1993)ConformationalepitopesofC3reflectingitsmodeofbindingtoanartificialpolymersurface.MolImmunol30:211–219.

Nissim A, Hoogenboom HR, Tomlinson IM, Flynn G, Midgley C, Lane D and Winter G. (1994)Antibody fragments from a 'single pot' phage display library as immunochemical reagents.EMBOJ13:692–698.

OhmuraN,TsukidateY,ShinozakiH,LackieSJandSaikiH. (2003)Combinationaluseofantibodyaffinities in an immunoassay for extension of dynamic range and detection of multipleanalytes.AnalChem75:104–110.

References

80

Ortega‐Vinuesa JL and Hidalgo‐Alvarez R. (1995) Sequential adsorption of F(ab')(2) and BSA onnegativelyandpositivelychargedpolystyrenelatexes.BiotechnolBioeng47:633–639.

OwenWE,RawlinsMLandRobertsWL. (2010)Selectedperformancecharacteristicsof therocheelecsys testosterone II assay on the modular analytics E 170 analyzer. Clin Chim Acta411:1073–1079.

OyamaH, Tanaka E, Kawanaka T,Morita I, Niwa T and Kobayashi N. (2013) Anti‐idiotype scFv–Enzymefusionproteins:Aclonableanalyte‐mimickingprobeforstandardizedimmunoassaystargetingsmallbiomarkers.AnalChem85:11553–11559.

PalamalaiV,MurakamiMandAppleF.(2013)Diagnosticperformanceoffourpointofcarecardiactroponin I assays to rule in and rule out acutemyocardial infarction.CliBiochem46:1631–1635.

PanteghiniM,GerhardtW,AppleFS,DatiF,RavkildeJandWuAH.(2001)Qualityspecificationsforcardiactroponinassays.ClinChemLabMed39:175–179.

Papapetrou PD, Polymeris A, Karga H and Vaiopoulos G. (2006) Heterophilic antibodies causingfalselyhighserumcalcitoninvalues.JEndocrinolInvest29:919–923.

Peluso P,WilsonDS, DoD, TranH:,M., QuincyD, Heidecker B, Poindexter K, Tolani N, PhelanM,WitteK,etal. (2003)Optimizing antibody immobilization strategies for the construction ofproteinmicroarrays.AnalChem312:113–124.

PernetP,Beneteau‐BurnatB,HermandCandVaubourdolleM. (2008)Point‐of‐care testing:Falseelevation of cardiac troponin I assayed in the emergency department. Am J Emerg Med26:969.e1–969.e2.

PerrySV.(1999)TroponinI:Inhibitororfacilitator.MolCellBiochem190:9–32.PiskinE,TuncelA,DenizliAandAyhanH.(1994)Monosizemicrobeadsbasedonpolystyreneand

their modified forms for some selected medical and biological applications. J Biomater SciPolymEd5:451–471.

PlebaniM,MionM,AltinierS,GirottoMA,BaldoGandZaninottoM.(2002)False‐positivetroponinIattributedtoamacrocomplex.ClinChem48:677–679.

PrestaLG.(2006)Engineeringoftherapeuticantibodiestominimizeimmunogenicityandoptimizefunction.AdvDrugDelivRev58:640–656.

Primus FJ, Kelley EA, Hansen HJ and Goldenberg DM. (1988) "Sandwich"‐type immunoassay ofcarcinoembryonic antigen in patients receivingmurinemonoclonal antibodies for diagnosisandtherapy.ClinChem34:261–264.

Prince AM, Brotman B, Jass D and Ikram H. (1973) Specificity of the direct solid‐phaseradioimmunoassayfordetectionofhepatitis‐Bantigen.Lancet1:1346–1350.

RaagRandWhitlowM.(1995)Single‐chainfvs.FASEBJ9:73–80.ReichlinT,HochholzerW,BassettiS,SteuerS,StelzigC,HartwigerS,BiedertS,SchaubN,BuergeC,

Potocki M, et al. (2009) Early diagnosis of myocardial infarction with sensitive cardiactroponinassays.NEnglJMed361:858–867.

Reinsberg J. (1995) Interference by human antibodies with tumor marker assays. Hybridoma2014:205–208.

Reinsberg J. (1996) Different efficacy of various blocking reagents to eliminate interferences byhumanantimouseantibodieswithatwo‐siteimmunoassay.ClinBiochem29:145–148.

Reinsberg J. (1998) Interferences with two‐site immunoassays by human anti‐mouse antibodiesformed by patients treated with monoclonal antibodies: Comparison of different blockingreagents.ClinChem44:1742–1744.

Renard M and Bedouelle H. (2004) Improving the sensitivity and dynamic range of reagentlessfluorescentimmunosensorsbyknowledge‐baseddesign.Biochemistry43:15453–15462.

RibattiD.(2014)Fromthediscoveryofmonoclonalantibodiestotheirtherapeuticapplication:Anhistoricalreappraisal.ImmunolLett161:96–99.

References

81

Ricchiuti V, Voss EM, Ney A, Odland M, Anderson PA and Apple FS. (1998) Cardiac troponin Tisoformsexpressed in renaldiseased skeletalmusclewillnot cause false‐positive resultsbythesecondgenerationcardiactroponinTassaybyboehringermannheim.ClinChem44:1919–1924.

Richman CM, DeNardo SJ, O'Grady LF and DeNardo GL. (1995) Radioimmunotherapy for breastcancer using escalating fractionated doses of 131I‐labeled chimeric L6 antibody withperipheralbloodprogenitorcelltransfusions.CancerRes55:5916s–5920s.

Ritter G, Cohen LS, Williams C,Jr, Richards EC, Old LJ andWelt S. (2001) Serological analysis ofhumananti‐humanantibodyresponsesincoloncancerpatientstreatedwithrepeateddosesofhumanizedmonoclonalantibodyA33.CancerRes61:6851–6859.

Robier C, Edler D and Neubar M. (2014) False‐positive D‐dimer result in a latex‐enhancedimmunoassay caused by interfering human anti‐mouse antibodies. Clin Chem Lab Med52:e253–e255.

RossHA,MenheerePP,EndocrinologySectionofSKML(DutchFoundationforQualityAssessmentinClinicalLaboratories),ThomasCM,MuddeAH,KouwenbergMandWolffenbuttelBH.(2008)Interference from heterophilic antibodies in seven current TSH assays. Ann Clin Biochem45:616.

Rotmensch S and Cole LA. (2000) False diagnosis and needless therapy of presumed malignantdisease inwomenwith false‐positive human chorionic gonadotropin concentrations.Lancet355:712–715.

Safdari Y, Farajnia S, Asgharzadeh M and Khalili M. (2013) Antibody humanization methods ‐ areviewandupdate.BiotechnolGenetEngRev29:175–189.

SampsonM,RuddelM, ZweigMandElinRJ. (1994)Falselyhigh concentrationof serum lutropinmeasuredwiththeabbottIMx.ClinChem40:1976–1977.

SandbornWJ, FeaganBG,Hanauer SB, PresentDH, Sutherland LR, KammMA,WolfDC, Baker JP,HawkeyC,ArchambaultA,etal.(2001)AnengineeredhumanantibodytoTNF(CDP571)foractive crohn'sdisease:A randomizeddouble‐blindplacebo‐controlled trial.Gastroenterology120:1330–1338.

SavukoskiT,EngströmE,EngblomJ,RistiniemiN,WittfoothS,LindahlB,EggersKM,VengePandPetterssonK.(2012)Troponin‐specificautoantibodyinterferenceindifferentcardiactroponinIassayconfigurations.ClinChem58:140–148.

Savukoski T, TwardaA, Hellberg S, RistiniemiN,Wittfooth S, Sinisalo J and PetterssonK. (2013)Epitope specificity and IgG subclass distribution of autoantibodies to cardiac troponin.ClinChem59:512–518.

Savukoski T, Ilva T, Lund J, Porela P, Ristiniemi N, Wittfooth S and Pettersson K. (2014)Autoantibody prevalence with an improved immunoassay for detecting cardiac troponin‐specificautoantibodies.ClinChemLabMed52:273–279.

SelbyC.(1999)Interferenceinimmunoassay.AnnClinBiochem36(Pt6):704–721.SgourisJT.(1973)LimitationsoftheradioimmunoassayforhepatitisBantigen(HBag).NEnglJMed

288:160–161.ShayanfarN,BestmannL,SchulthessGandHersbergerM.(2008)False‐positivecardiactroponinT

duetoassayinterferencewithheterophilicantibodies.SwissMedWkly138:470.ShimbaN,TorigoeH,TakahashiH,MasudaK,ShimadaI,ArataYandSaraiA.(1995)Comparative

thermodynamicanalysesof the fv, fab*and fab fragmentsof anti‐dansylmousemonoclonalantibody.FEBSLett360:247–250.

Shirahama H and Suzawa T. (1984) Surface characterization of soap‐free carboxylated polymerlatices.PolymerJournal16:795–803.

Shmilovich H, Danon A, Binah O, Roth A, Chen G, Wexler D, Keren G and George J. (2007)Autoantibodies to cardiac troponin I in patients with idiopathic dilated and ischemiccardiomyopathy.IntJCardiol117:198–203.

References

82

Simpson AJ, Potter JM, Koerbin G, Oakman C, Cullen L, Wilkes GJ, Scanlan SL, ParsonageW andHickmanPE.(2014)Useofobservedwithin‐personvariationofcardiactroponininemergencydepartment patients for determination of biological variation and percentage and absolutereferencechangevalues.ClinChem60:848–854.

Sodi R, Darn SM, Davison AS, Stott A and Shenkin A. (2006) Mechanism of interference byhaemolysisinthecardiactroponinTimmunoassay.AnnClinBiochem43:49–56.

SoiniEandKojolaH.(1983)Time‐resolvedfluorometerforlanthanidechelates–anewgenerationofnonisotopicimmunoassays.ClinChem29:65–68.

Sosolik RC, Hitchcock CL and Becker WJ. (1997) Heterophilic antibodies produce spuriouslyelevated concentrations of the MB isoenzyme of creatine kinase in a selected patientpopulation.AmJClinPathol107:506–510.

Soukka T, Härmä H, Paukkunen J and Lövgren T. (2001a) Utilization of kinetically enhancedmonovalent binding affinity by immunoassays based on multivalent nanoparticle‐antibodybioconjugates.AnalChem73:2254–2260.

SoukkaT,Paukkunen J,HärmäH,LönnbergS,LindroosHandLövgrenT. (2001b)Supersensitivetime‐resolved immunofluorometric assay of free prostate‐specific antigenwith nanoparticlelabeltechnology.ClinChem47:1269–1278.

Soukka T, Antonen K, Härmä H, Pelkkikangas AM, Huhtinen P and Lövgren T. (2003) Highlysensitive immunoassay of free prostate‐specific antigen in serum using europium(III)nanoparticlelabeltechnology.ClinChimActa328:45–58.

St John A and Price CP. (2013) Economic evidence and point‐of‐care testing. Clin Biochem Rev34:61–74.

Staros JV, Wright RW and Swingle DM. (1986) Enhancement by N‐hydroxysulfosuccinimide ofwater‐solublecarbodiimide‐mediatedcouplingreactions.AnalBiochem156:220–222.

Steimer W. (1999) Performance and specificity of monoclonal immunoassays for cyclosporinemonitoring:Howspecificisspecific?ClinChem45:371–381.

Storrow AB, Apple FS, Wu AHB, Jesse RL, Francis GS, Christenson RH, Cannon C,P, Morrow DA,Newby LK, Ravkilde J, et al. (2007) National academy of clinical biochemistry laboratorymedicine practice guidelines: Point of care testing, oversight, and administration of cardiacbiomarkersforacutecoronarysyndromes.PointCareJ6:215–222.

Storrow AB, Zhou C, Gaddis G, Han JH, Miller K, Klubert D, Laidig A and Aronsky D. (2008)Decreasing lab turnaroundtime improvesemergencydepartmentthroughputanddecreasesemergencymedicalservicesdiversion:Asimulationmodel.AcadEmergMed15:1130–1135.

Sturgeon CM and McAllister EJ. (1998) Analysis of hCG: Clinical applications and assayrequirements.AnnClinBiochem35(Pt4):460–491.

SturgeonCMandViljoenA. (2011)Analyticalerrorand interference in immunoassay:Minimizingrisk.AnnClinBiochem48:418–432.

Tamaki K and Shimomura M. (2002) Fabrications of luminescent polymeric nanoparticlescontaininglanthanide(III)ioncomplexes.IntJNanosci1:533–537.

TateJRandWardG.(2004)Interferencesinimmunoassay.ClinBiochemRev25:105–120.TateJR.(2008)Troponinrevisited2008:Assayperformance.ClinChemLabMed46:1489–1500.ThomasCMandSegersMF.(1985)Discordantresultsforchoriogonadotropin:Aproblemcausedby

lutropinbeta‐subunitinterference?ClinChem31:159.Thompson RJ, Jackson AP and Langlois N. (1986) Circulating antibodies to mouse monoclonal

immunoglobulins innormal subjects‐‐incidence, speciesspecificity, andeffectsona two‐siteassayforcreatinekinase‐MBisoenzyme.ClinChem32:476–481.

Thygesen K, Alpert JS, Jaffe AS, SimoonsML, Chaitman BR,White HD and theWriting Group onbehalfoftheJointESC/ACCF/AHA/WHFTaskForcefortheUniversalDefinitionofMyocardialInfarction. (2012)Thirduniversaldefinitionofmyocardial infarction.Circulation126:2020–2035.

References

83

ToddJ,FreeseB,LuA,HeldD,MoreyJ,LivingstonRandGoixP.(2007)Ultrasensitiveflow‐basedimmunoassaysusingsingle‐moleculecounting.ClinChem53:1990–1995.

TurpeinenU,LehtovirtaP,AlfthanHandStenmanUH. (1990) Interferencebyhumananti‐mouseantibodies in CA 125 assay after immunoscintigraphy: Anti‐idiotypic antibodies notneutralizedbymouseIgGbutremovedbychromatography.ClinChem36:1333–1338.

Uettwiller‐Geiger D, Wu AH, Apple FS, Jevans AW, Venge P, Olson MD, Darte C, Woodrum DL,RobertsSandChanS.(2002)MulticenterevaluationofanautomatedassayfortroponinI.ClinChem48:869–876.

VaidyaHCandBeattyBG.(1992)Eliminatinginterferencefromheterophilicantibodiesinatwo‐siteimmunoassay for creatine kinaseMBby using F(ab')2 conjugate and polyclonalmouse IgG.ClinChem38:1737–1742.

Väisänen V, Peltola MT, Lilja H, Nurmi M and Pettersson K. (2006) Intact free prostate‐specificantigenandfreeandtotalhumanglandularkallikrein2.eliminationofassayinterferencebyenzymaticdigestionofantibodiestoF(ab')2fragments.AnalChem78:7809–7815.

Valanne A, Huopalahti S, Soukka T, Vainionpää R, Lövgren T and Härmä H. (2005) A sensitiveadenovirus immunoassay as a model for using nanoparticle label technology in virusdiagnostics.JClinVirol33:217–223.

vanGageldonkPG,vonHunolsteinC,vanderKlisFRandBerbersGA.(2011)Improvedspecificityofamultipleximmunoassayforquantitationofanti‐diphtheriatoxinantibodieswiththeuseofdiphtheriatoxoid.ClinVaccineImmunol18:1183–1186.

VengeP,JohnstonN,LagerqvistB,WallentinL,LindahlPandFRISC‐IIStudyGroup.(2003)ClinicalandanalyticalperformanceoftheliaisoncardiactroponinIassayinunstablecoronaryarterydisease,andthe impactofageonthedefinitionofreference limits.AFRISC‐IIsubstudy.ClinChem49:880–886.

Venge P, Johnston N, Lindahl B and James S. (2009) Normal plasma levels of cardiac troponin Imeasuredbythehigh‐sensitivitycardiactroponinIaccessprototypeassayandtheimpactonthediagnosisofmyocardialischemia.JAmCollCardiol54:1165–1172.

ver Elst KM, Chapelle JP, Boland P, Demolder JS and Gorus FK. (1999) Analytic and clinicalevaluationoftheabbottAxSYMcardiactroponinIassay.AmJClinPathol112:745–752.

vonLodeP,HagrenV,PaleniusTandLövgrenT.(2003)One‐stepquantitativethyrotropinassayforthedetectionofhypothyroidisminpoint‐of‐careconditions.ClinBiochem36:121–128.

Vylegzhanina AV, Katrukha IA, Kogan AE and Bereznikova AV. (2013) Epitope specificity of anti‐cardiactroponinImonoclonalantibody8I‐7.ClinChem59:1814–1816.

Wakayama J, Sekiguchi H, Akanuma S, Ohtani T and Sugiyama S. (2008) Methods for reducingnonspecific interaction inantibody‐antigenassayviaatomic forcemicroscopy.AnalBiochem380:51–58.

Walker JBandNesheimME. (1999)Themolecularweights,massdistribution, chain composition,andstructureofsolublefibrindegradationproductsreleasedfromafibrinclotperfusedwithplasmin.JBiolChem274:5201–5212.

WangY,DosseyAM,FroudeJW,2nd,LubitzS,TzurD,SemenchenkoVandWishartDS.(2008)PSAfluoroimmunoassaysusinganti‐PSAScFvandquantum‐dotconjugates.Nanomedicine(Lond)3:475–483.

WardG,McKinnonL,BadrickTandHickmanPE.(1997)Heterophilicantibodiesremainaproblemfortheimmunoassaylaboratory.AmJClinPathol108:417–421.

WarrenDJ,BjernerJ,PausE,BørmerOPandNustadK.(2005)Useofaninvivobiotinylatedsingle‐chain antibody as capture reagent in an immunometric assay to decrease the incidence ofinterferencefromheterophilicantibodies.ClinChem51:830–838.

Waser G, Kathriner S and Wuillemin WA. (2005) Performance of the automated and rapid STAliatestD‐dimerontheSTA‐Ranalyzer.ThrombRes116:165‐170.

Waterboer T, Sehr P and Pawlita M. (2006) Suppression of non‐specific binding in serologicalluminexassays.JImmunolMethods309:200–204.

References

84

WegenerWA,PetrelliN,SerafiniAandGoldenbergDM.(2000)Safetyandefficacyofarcitumomabimagingincolorectalcancerafterrepeatedadministration.JNuclMed41:1016–1020.

WilsonDS,WuJ,PelusoPandNockS. (2002) Improvedmethod forpepsinolysisofmouse IgG(1)moleculestoF(ab')(2)fragments.JImmunolMethods260:29–36.

Wolfe F, CatheyMAandRoberts FK. (1991)The latex test revisited. rheumatoid factor testing in8,287rheumaticdiseasepatients.ArthritisRheum34:951–960.

WuAH, Apple FS, GiblerWB, JesseRL,WarshawMMandValdesRJ. (1999)National academyofclinical biochemistry standards of laboratory practice: Recommendations for the use ofcardiacmarkersincoronaryarterydiseases.ClinChem45:1104–1121.

WuAH,AgeeSJ,LuQA,ToddJandJaffeAS.(2009)Specificityofahigh‐sensitivitycardiactroponinIassayusingsingle‐molecule‐countingtechnology.ClinChem55:196–198.

WuAH,Akhigbe P andWians F. (2012) Long‐termbiological variation in cardiac troponin I.ClinBiochem45:714–716.

WuAHandChristensonRH.(2013)Analyticalandassayissuesforuseofcardiactroponintestingforriskstratificationinprimarycare.ClinBiochem46:969–978.

XuL,YuZ,FanW,WangX,XieM,XuY,HuLandLiY.(2013)NegativeinterferenceinserumHBsAgELISAfromrheumatoidfactors.PLoSOne8:e80620.

Yalow RS and Berson SA. (1959) Assay of plasma insulin in human subjects by immunologicalmethods.Nature184(Suppl21):1648–1649.

YeoKT,StormCA,LiY,JayneJE,BroughT,Quinn‐HallKSandFitzmauriceTF.(2000)Performanceof the enhanced abbott AxSYM cardiac troponin I reagent in patients with heterophilicantibodies.ClinChimActa292:13–23.

Ylikotila J, Hellström JL, Eriksson S, Vehniäinen M, Välimaa L, Takalo H, Bereznikova A andPettersson K. (2006) Utilization of recombinant fab fragments in a cTnI immunoassayconductedinspotwells.ClinBiochem39:843–850.

ZaidiAandCowellR.(2010)Falsepositivecardiactroponinelevationduetoheterophileantibodies:Morecommonthanwerecognise?BMJCaseRep2010:1–3.

ZhangX,XuC,ZhangC,LiuY,XieYandLiuX.(2014)Establishedanewdoubleantibodiessandwichenzyme‐linked immunosorbent assay for detecting bacillus thuringiensis (bt) Cry1Ab toxinbased single‐chain variable fragments from a naïvemouse phage displayed library.Toxicon81:13–22.

Zhou L, Ding F, Chen H, DingW, ZhangW and Chou SY. (2012) Enhancement of immunoassay’sfluorescenceanddetectionsensitivityusingthree‐dimensionalplasmonicnano‐antenna‐dotsarray.AnalChem84:4489–4495.

ZhuY, JenkinsMM,BrassDA,RavagoPG,HorneBD,DeanSBandDraytonN. (2008)Heterophilicantibodyinterferenceinanultra‐sensitive3‐sitesandwichtroponinIimmunoassay.ClinChimActa395:181–182.

Zisman A, Zisman E, Lindner A, Velikanov S, Siegel YI and Mozes E. (1995) Autoantibodies toprostatespecificantigeninpatientswithbenignprostatichyperplasia.JUrol154:1052–1055.

Date post:	26-Jun-2020
Category:	Documents
Upload:	others
View:	0 times
Download:	0 times

NANOPARTICLE-ASSISTED IMMUNOASSAYS FOR POINT-OF … · 2017-02-22 · reductions in the assay...

Documents