Residence time distribution in counter-current protein A affinity chromatography using an inert tracer

The trend in the biopharmaceutical industry is changing from batch process to continuous process. For continuous biomanufacturing, traceability of the material is required by regulatory authorities.

Journal of Chromatography A 1683 (2022) 463530

ContentslistsavailableatScienceDirect

journalhomepage:www.elsevier.com/locate/chroma

Narges Lalia,b, Peter Satzera,b, Alois Jungbauera,b,∗

a ACIB- Austrian Centre of Industrial Biotechnology, Krenngasse 37, Graz 8010, Austria

b Department of Biotechnology, University of Natural Resources and Life Sciences, Vienna, Austria

Article history:

Received 1 July 2022

Revised 19 September 2022

Accepted 21 September 2022

Available online 22 September 2022

Keywords:

Residence time distribution

Chromatography

Staphylococcal protein A

Periodic counter-current chromatography

Continuous

a b s t r a c t

The trend in the biopharmaceutical industry is changing from batch process to continuous process For continuous biomanufacturing, traceability of the material is required by regulatory authorities The recent ICH draft guideline Q13 on continuous manufacturing of drug substances and drug products requests an

“understanding of process dynamics as a function of input material attributes (e.g., potency, material flow properties), process conditions (e.g., mass flow rates) … One common approach is characterization

of residence time distribution (RTD) for the individual unit operations and integrated system.” Thus, it

is necessary to trace material through individual continuous unit operations and the integrated process The RTD of a process is obtained experimentally by injecting a pulse of an inert tracer into the inlet and measuring the broadening of the injected pulse in the outlet We investigated the RTD of three-column periodic counter-current chromatography (PCC) using staphylococcal protein A affinity chromatography, with a focus on how the material distributes over subsequent cycles A fluorescent-labeled antibody was used as the inert tracer under high salt concentration The tracer was injected once in each run but at different points of the loading phase We then analyzed the outlet of the column In the elution phase, regardless of the point of injection, we observed an even distribution of the tracer In the loading phase, a constant exchange between the antibody in the solid phase and the liquid phase was observed, meaning that sending the outlet of one chromatography column into another column to improve resin utilization causes higher residence time in the system for some portion of the material

© 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/)

1 Introduction

The biopharmaceutical industry has recently become highly

interested in moving from batch process to continuous process,

though there are many challenges along the way, including the

traceabilityofmaterialthroughtheprocess.Traceabilityofthe

ma-terialisrequiredbyregulatoryauthoritiesforcontinuous

bioman-ufacturing.TherecentICHdraftguidelineQ13oncontinuous

man-ufacturingofdrugsubstancesanddrugproductsasksforthe

char-acterization of material flow in the continuous process and

rec-ommendsusingtheresidencetimedistribution(RTD)forthis

pur-pose The RTDis the probability distribution ofthe time a piece

of a substance is likely to spend in the reactor The most

com-monmethodofdeterminingtheRTDistomeasurethesystem

re-sponse foraninerttracer pulse[1].Theknowledge ofthis

distri-bution isimportantwhendesigning theprocess,aswell aswhen

∗ Corresponding author at: ACIB- Austrian Centre of Industrial Biotechnology,

Krenngasse 37, Graz 8010, Austria

E-mail address: alois.jungbauer@boku.ac.at (A Jungbauer)

theprocess isperformed, in ordertoget informationonstart-up andshut-down timing and samplingfrequency to determine the adequateprocessanalyticaltechnology(PAT)[2].Furthermore,the challengeofdefininganewbatchdefinitionforacontinuous bio-processwasaddressedbyRTDcharacterization[3,4].RTDneedsto

bedefinedfortheindividual unitoperationsandthenforthe in-tegratedprocess

RTDis measured through a tracer experiment, onlineprocess measurements of an appropriate product attribute, and/or pro-cessmodeling[3,5–8].RTDiscalculatedtheoreticallyusingamass transferequation,thoughinchromatographyitisquitechallenging becausethereareadsorptionanddesorptionkineticsinvolved Cur-rentunderstandingoftheadsorptionofantibodiesonporous pro-teinAaffinitychromatographybeadsisaporediffusion-controlled process[9–14].Thesaturationoftheporousbeadcanbedescribed

asashrinking coremodelinwhichthe saturatingmigratedfrom theoutersurfaceofthebeadtothecore[15].Forsimplerunit op-erations, such ascontinuously stirring tank (CSTR) and plug flow reactor, the RTD function is fully solved for ideal cases [16] For morecomplicatedunit operationsorprocesses,RTDcan be

mea-https://doi.org/10.1016/j.chroma.2022.463530

0021-9673/© 2022 The Author(s) Published by Elsevier B.V This is an open access article under the CC BY license ( http://creativecommons.org/licenses/by/4.0/ )

Fig 1 Schematic diagram of three-column periodic counter-current chromatography (PCC) in one cycle including six steps Steps 2, 4, and 6 show the main loading steps in

which two columns are in the loading phase and the other column is in the elution/regeneration step In these steps, the breakthrough of a column goes to the next column

to increase resin utilization Steps 1, 3, and 5 show the wash phase between the main loading steps

suredexperimentally by injectinga smallpulseofaninerttracer

andthentracingitintheoutlet.Theinerttracerneedstobe

iden-tical to the proteinof interest, butwe also need a difference by

which to detect the tracer Thus, finding an inerttracer is

chal-lenging

Antibodypurificationisperformedconventionallyby proteinA

affinity chromatography in batch mode; however, increasing

de-mandforantibodyproductslettheindustrymovefrombatch

pro-cessingtocontinuousprocessing.Abatchchromatographyprocess

isconvertedintoacontinuousprocessbyapplyingcounter-current

loading, also known as periodic counter-current chromatography

(PCC)[17].Thecombinationoffed-batchandcontinuouscaptureis

the mosteconomicalwaytoproduce monoclonalantibodies[18]

Thisshiftfrombatch tocontinuousprocessingleadstohigher

ca-pacity,higherquality,andlowerprices,buttheimprovedprocess

economicsisobtainedattheexpenseofcomplexity[18].TheRTD

of batch chromatographyislesschallengingbecause thematerial

is not transferred from one column to another, and the elution

is pooled and stirred Furthermore, we assume that the column

is fullyregenerated and the antibody is not left on the column

We assume a clean columnforevery batch Yet,in PCC ormany

continuous chromatography unit operations, part of the material

is transferredfromonecolumntothe nextcolumn, soitis more

challengingtotracethematerial

Labeling the monoclonalantibody with a fluorescent dye and

usingthefluorescentsignalfordetectionisamethodthathasbeen

widelyusedinchromatography,mainlyforconfocallaserscanning

microscopy(CLSM)imaging.Thequestionaroseastowhetherthe

labeled antibody has the exact properties of the unlabeled

anti-body because a slight difference may result in different binding

propertiesand/or displacementandthedangerofmisinterpreting

experimental results [19–22] Quantum dots may have a higher

sensitivity, butthereis lessexperience withthechromatographic

behavior ofsuch labeledproteins,anditisobviousthatthey also

introduceanaberrantchargetotheprotein[23].Despitethe

criti-cal aspectsoffluorescentlylabelingantibodies,themethodhasits

advantages: It isvery sensitive, anda variety offluorescent dyes

are readily available [24].This methodis cost-effectivecompared

to radiolabeling, and on-line scintillation counters are not

com-monlyusedinbioprocessinglaboratories

The goal ofthis work isto characterize the RTD ofprotein A chromatographyintheloadandelutionphases.First,we setup a methodtodemonstratethat thefluorescentlylabeled monoclonal antibody serves as an inerttracer, and then we determined the RTDofacounter-currentproteinAaffinitychromatographyrunby injectingthetraceratthestart,middle,andendofloading

Acommercialmonoclonalantibody(IgG2)wasusedasthe pro-tein ofinterest.Forthelast setofexperiments, another commer-cialmonoclonalantibody(IgG1)wasused.Chemicalsandreagents were purchasedfromMerck KGaAorSigmaAldrichunless other-wise stated All bufferswere filtered by a 0.22 μm filter(Merck KGaA)beforeuse.Afluorescentproteinlabelingkit(AlexaFluorTM 488)waspurchasedfromInvitrogen,anda10kDamembranefilter (MilliporeAmicon)wasusedduringthelabelingprocess

Fluorescent dyewas used to label the antibody The free dye wasremovedusinga10kDamembranefilter.Thedegreeof label-ing(DOL)wasdeterminedaccordingtothemanufacturer’s instruc-tions.TheDOLwasalways<1toensureminimallabeling.The fi-nalmonoclonalantibodylabeledwiththefluorescentdyewas re-ferredtoasthelabeledantibody

AnHPLCsystem(DionexUltimate3000HPLCsystem)and an-alytical protein A affinity chromatography column were used to determine the antibody concentration and fluorescence intensity The HPLC systemwas equippedwith a diode array detectorand fluorescent detector(ThermoFisher Scientific) The analytical col-umnwasaPOROSA20μmcolumn(2.1× 30mm,0.1ml; Ther-moFisher Scientific) The equilibration buffer was 50 mM phos-phatebuffercontaining150mMNaCl(pH7.0).Theelution buffer was100mMglycinebuffer(pH2.5).Thesampleinjectionwas20

N Lali, P Satzer and A Jungbauer Journal of Chromatography A 1683 (2022) 463530

Fig 2 Batch mode condition equivalent to one column in counter-current chro-

matography (PCC) The blue area indicates the product bound to resin in a column

The brown area indicates the unbound product leaving the column a) Full break-

through curve The blue area shows the maximum resin capacity, which is equal to

the equilibrium binding capacity (EBC) b) Conventional load volume in batch mode

until 10% of breakthrough (DBC10) c) The batch mode loading volume is equivalent

to the PCC loading volume, and the load is up to 70% of breakthrough d) The batch

mode chromatography up to 70% breakthrough, which reproduces the RTD in one

column in PCC The dashed line shows the switches between different steps in one

cycle For example, column 2 from Fig 1 : Section 1 shows the material bound to

column 2 from column 3 in the previous wash step (step 1 in Fig 1 ), and Section

2 shows the material bound to column 2 from the breakthrough of column 1 (step

2 in Fig 1 ), Section 3 shows the material bound to column 2 in the main loading

step (step 3 in Fig 1 ), Section 4 shows the material moved out of column 2 and

loaded into column 3 (step 4 in Fig 1 ), and Section 5 shows the material moved

out of column 2 and loaded into column 1 (step 5 in Fig 1 ) Area V is equal to area

I and area IV is equal to area II

μl The UV signal at 280 nm (UV280) was monitored and used

to determinetheantibody concentration,andthefluorescent

sig-nal was monitored and used to determine the labeled antibody

concentration The results were evaluated and quantified using

Chromeleon7software(ThermoFisherScientific)

For continuous chromatography, three-column PCC was used

(Fig.1) OnePCCcycleconsistsofsixsteps Steps2,4,and6are

the mainloading steps,inwhichtwo columnsare intheloading

phase and one columnisin the elution/regeneration phase

Dur-ing thesesteps,theinlet goestoa columnandthebreakthrough

of that column goes to the next column, so resin utilization in-creases.There is a UV detector betweenthe two columnsin the loadingphase tomonitorthe breakthroughcurve Steps1,3,and

5 are thewash steps betweenthe main loadingsteps The wash buffergoestothefullyloaded columnandsendsthewashbuffer together with the remaining product to a freshcolumn to avoid losingtheproduct.Thesestepsareonlyaslongasthewashphase aftertheloadingphase(Fig.1)

The system was an ÄKTA Avant chromatography system equippedwithextravalvesandasecondUVdetector.Extratubing wasaddedwhennecessary,andtheflowpathwaschangedto pro-duceathree-columnPCCsystemasexplainedbyGomis-Fonsetal [20].Mab SelectSuRe(Cytiva)resin waspackedintoTricorn5/50 columns,andthecolumnvolumeforthethreecolumnswas1ml The equilibration/load/wash buffer was20 mM phosphatebuffer containing150mMNaCl(pH7.4).Theelutionbufferwas100mM acetate buffer (pH3.0) For some experiments, we increased the NaClto500mMintheequilibration/load/washbufferandelution buffer Chromatographyphasesincludedequilibration,load, wash, andelutionphases.Theflowratewas0.4ml/min

In the PCC experiments, loading time/volume was calculated basedona preliminaryexperimentandmonitoredby thesecond

UV detector,which wasplaced afterthefirstcolumninthemain loadingphases[25].Loading time/volumewasbased onreaching 70% ofthe final breakthrough signal [25].The UV280 was moni-toredduringtheelutionphase.FortheRTDmeasurement,0.5ml

oflabeledantibodywasinjectedduringtheloadingphaseusinga sampleloopintheinjectionvalve.Allelutionpeakswerecollected

in0.5mlor1mlfractionsandanalyzedforantibodyconcentration andlabeledantibodyconcentration

The basic principle of many continuous chromatography unit operationsis to increase the loadamount to maximize the resin utilizationby recyclingtheoutletofthecolumntoavoid decreas-ing the yield [17] We used a batch chromatography experiment

toreproducetheresultsfromonlyonecolumninPCC,which ap-pliesto differentcontinuouschromatographyprocesses,including PCC.Thebatchchromatographywassetupbyprolongingtheload phase up to 70% of the final breakthrough (Fig 2) Figs 1 and

2showthat in onecolumn in PCC,bound material camefrom the two previous steps (andcolumn), andtheunboundproduct goes intothenexttwosteps(andcolumn);therefore,itmakestheRTD

inPCCrathercomplicated

Theprolonged loadingphase inbatch chromatographyisused

asanequivalenttocharacterizethe PCC(Fig.2) Forexample,for column 2 (Figs 1 and 2), there are different sections of loaded material: Section 1 shows the material bound to column2 from column3 in the previous washstep (step), Section 2 shows the material bound to column2from the breakthroughof column1 (step 2),Section 3 showsthematerial bound tocolumn2 inthe main loading step (step 3), Section 4 showsthe material moved outofcolumn2andloadedintocolumn3(step4),andSection5

showsthe material moved out ofcolumn 2andloaded into col-umn1(step 5).Bytracing thematerialinthesesections,we pre-dictthematerialflowinathree-columnPCC.Theadvantageofthis experimentisthatfractionationoftheoutletofonecolumninthe loadingphaseispossible

The same ÄKTA Avant chromatography system was used but

inthe conventionalflow pathequipped withoneofthe columns mentionedin2.4.basedonreaching70%ofthefinalbreakthrough signal.UV280wasmonitoredduringtheelutionphase.FortheRTD measurement,0.5mloflabeledantibodywasinjectedduringthe loadingphaseusingasampleloopintheinjectionvalve.The

out-Fig 3 Two elution peaks in PCC Left, with conventional protein A buffer Right, with 500 mM of all buffers Blue bars show the unlabeled antibody concentration Green

bars show the labeled antibody concentration In the elution peak with conventional buffers, the retention time of the labeled antibody is longer than that of the unlabeled antibody In the elution peak with high salt buffer, the retention time of the labeled antibody is equal to that of the unlabeled antibody

Fig 4 Three consecutive elution peaks in a PCC experiment Blue bars show the unlabeled antibody concentration Green bars show the labeled antibody concentration The

labeled antibody was injected at the start of the loading phase corresponding to the first elution peak, but it was detectable in the next two elution peaks after injection

letofallphaseswascollectedin0.5mlor1mlfractionsand

an-alyzed byanalytical HPLC forantibodyconcentration andlabeled

antibodyconcentration

A Leica TCS SP8-STED laser (point) scanning confocal

micro-scope wasused Fluorescentlylabeled antibodywasused for

im-age acquisition (green) The microscope was equipped with the

StimulatedEmissionDepletion(STED)module,whichoffersspatial

super-resolution fluorescence imaging The electronics and

soft-waremodulewere fromPicoQuant(Germany).ForAlexa488dye,

the excitation wavelength was490 nm andemission wavelength

520nm

Achannelslide(ibidiGmbH)withaheightof0.2mmwas con-nectedtoasyringepump,onefilterfromaTricorn5filterkit (Cy-tiva)wasplaced intotheoutletoftheLuer connection,andresin beadswerelooselypackedintothechannel,producinga microcol-umn.Thesyringepumpwasusedtorunthemicrocolumnwiththe samesuperficial velocity.Themicrocolumn wasplaced underthe microscopeandan image wasdetected every 10s.A chromatog-raphy run was performedunder the microscope, including equi-libration, loading (unlabeled) antibody, loading labeled antibody, loading(unlabeled) antibody,wash,andelution Duringthe load-ing,theunlabeledantibody,equalto40%oftheresincapacity,was loadedintotheresinfirst.Next,thelabeledantibody,equalto40%

oftheresincapacity,andtheunlabeledantibody, equalto40%of

N Lali, P Satzer and A Jungbauer Journal of Chromatography A 1683 (2022) 463530

Fig 5 Injection of labeled antibody at the end of the loading phase The blue line shows the unlabeled antibody concentration The green line shows the labeled antibody

concentration

resincapacity,wereloaded.Intotal,theamountwasequalto120%

oftheresincapacity

3 Result and discussion

During the PCC experiments with three protein A affinity

columns,0.5mloffluorescent-labeledantibodywasinjectedeither

atthestart,themiddle,ortheendoftheloadingphase.The

un-labeled and labeledantibody concentrationsin the elution peaks

show that, when conventional buffers (lower salt concentration)

were used,regardlessofthepointofinjection,the average

reten-tion time ofthelabeled antibodieswaslonger thanthat of

unla-beledantibody(Fig.3).Thissuggeststhat,underthiscondition,the

affinity ofthelabeled antibodywashigherthanthatofunlabeled

antibody.Inaddition,thismeansthatthefluorescentlylabeled

an-tibodydoesnotactasaninerttracer.Thefluorescentdyewas

pre-viouslydemonstratedtoaltertheretentiontimeofthelabeled

an-tibody becauseitleadstoslightdifferencesinhydrophobicityand

charge However, thisdifferenceinretentiontime waseliminated

byincreasingthesaltconcentration[21]

Thus, the PCC experiments were performed again with

in-creased salt concentration: 500 mM NaCl in the

equilibra-tion/load/wash and elution buffer Under highsalt concentration,

regardless ofthepoint ofinjection, theaverage retentiontime of

the labeledantibodywasequalto thatofthe unlabeledantibody

(Fig.3).Therefore,weconcludethatthelabeledantibodywas

con-sidered an inert tracer under high salt concentration conditions

In addition, theresults indicatethat any section ofthe inlet

dis-tributesequallyintheelutionpeak.Forthisreason,allfurther

ex-perimentswereperformedunderhighsaltconcentrations

By increasing the NaCl concentration in the equilibra-tion/load/washandelutionbufferto500mM,wehadthelabeled antibodyasaninerttracer.However,thelabeledantibodywas un-expectedlydetectedinthenextelutionpeaks(Fig.4),whichmeans thatthelabeledmaterialthatwasalreadyloadedintoonecolumn wastransferredduringtheloadingphasetothenextcolumn Fur-therinvestigationofthiseffectrequiredfractionationandanalysis

oftheoutletofonecolumninPCC.Thelimitationwasthat,inthe PCC process, the outlet of one column goes directly to the next column,socollecting theoutletofonecolumn(Fig.1,step2, be-tween column1 andcolumn2) interrupts the wholeprocess To overcomethischallenge,a batchchromatographyexperimentwas setupwiththeconditionsmentionedin2.5torepresenttheRTD

ofonecolumninthePCC

Asexplainedin2.5,batch chromatographywithalonger load-ing volume can be representative of the RTD of each column in continuouschromatography (Fig 2) Batchmode chromatography wasperformed to better understand the RTDduring the loading phase.Thepulseinjectionwasperformedduringtheloadingphase butatdifferentpointsoftheloadingphase

Injectionattheendoftheloadingphase showedthatthe an-tibody bound to the resin more than expected When the tracer wasinjected atthe point of60% breakthrough,we expected 60%

oflabeledantibodytoleavetheresinand40%tobindtotheresin; thus,this40%wouldbefoundintheelutionpeak However,only 20%ofinjectedlabeledantibodymovedoutofthecolumnduring theloadingphaseand80%wasintheelutionpeak(Fig.5)

Fig 6 Injection of labeled antibody at the start of the loading phase The blue line shows the unlabeled antibody concentration The green line shows the labeled antibody

concentration

Fig 7 Injection of labeled antibody at the start of the loading phase, fully overloading the column until reaching a full breakthrough and continuing to load until 220 CV

The blue line shows the unlabeled antibody concentration The green line shows the labeled antibody concentration

Whenthetracerwasinjectedatthestartoftheloadingphase,

itstartedleavingthecolumntogetherwithanexcessofunlabeled

antibodyinthebreakthrough.Theinitialassumptionwasthatthe

tracer isbound totheresinuntiltheelutionphase.However,10%

oftheinjectedmaterialmovedoutofthecolumnduringthe

load-ingandwashphase(Fig.6)

HIC analytical chromatographyand CEX analytical chromatog-raphywereperformed(datanotshown)toidentifythedifferences between the labeled and unlabeledantibodies The labeled anti-bodyisslightlymorehydrophobic.AntibodyproteinAinteraction

isdominatedbyhydrophobicinteractions[26,27].Thus,amore

hy-N Lali, P Satzer and A Jungbauer Journal of Chromatography A 1683 (2022) 463530

drophobic tracer binds stronger, but we observed that a certain

fractionexitsthecolumneventhoughitshouldremaininside

Weconcludethatthedesorptionreactionbecomesmore

signif-icantespeciallywhentheresinisoverloadedandtheconcentration

ofantibodyishighinthesolidphase.Weassumethatconstant

ex-changebetweentheantibodyinthe liquidphaseandsolidphase

isoccurring.Thedesorptionisdetectabletoalesserextentin

con-ventionalbatchchromatographywhentheloadingvolumeisupto

10%ofbreakthrough.Inthatcase,theresinisnotfullyloaded,and

the desorbed product getsadsorbed againonto the unused fresh

resinfurtherinthesamecolumn Althoughthedesorptionratein

batchchromatographyisusuallyconsiderednegligible,previous

re-searchhasshownthattheoveralladsorptionrateisacombination

ofthe adsorption/desorptionrate[28].The loadamount ishigher

ineachcolumninPCC,sothedesorptionrateismorepronounced

whenitcomestocontinuouschromatography

Following the pulseinjection in batch chromatography

(injec-tion atthestartofthe loadingphase), anexperimentwassetup

to investigateif the tracer can be completely removed fromthe

column by prolonging the loading phase andadding more

unla-beled antibody The assumption wasthat, if we loadmore

unla-beledantibody,theconstantexchangebetweentheliquidandsolid

phasewouldleadtolosingthelabeledantibodyduringtheloading

phase.Weloadedacolumnuntilfull breakthroughandcontinued

foranexcessivelylongloadingvolume(220CV).Duringthe

load-ing phase, 65% of the tracer moved out of the column, andthe

concentrationoflabeledantibodyintheflowthroughdecreasedas

theoverallconcentrationinthecolumndecreased(Fig.7).This

ex-perimentconfirmedthatthereisaconstantexchangebetween

an-tibodiesintheliquidandsolidphases

To confirmthis unexpectedexchange, CLSMimaging was

per-formed tovisualizetheeffect(Fig.8) CSLMhasbeenused

previ-ouslywhentheexacttransportmechanismisstillunresolved[29]

Thebeadswereplacedinthemicrochannelandthe

chromatogra-phyrun performedunderthemicroscope, includingequilibration,

loadingunlabeledantibody,loadinglabeledantibody,loading

unla-beled antibody,wash,andelution(seeSection 2.6).Thebeadsare

loaded withunlabeledantibodyuntil 40%ofthebinding capacity

isreached Theunlabeledantibodyisnotvisible(Fig.8B) The

la-beled antibody,equalto40%ofthebindingcapacity,wasthen

in-jected,followedbyloadingtheunlabeledantibody,equalto40%of

thebinding capacity(Fig.8D).TheCLMSimagesshowthat,when

thelabeledantibodywasadded,itstartedtobindfromthesurface

of theresin even though the antibodywasalready bound tothe

resin.Thisbehaviorisexpectedforaporediffusion-limitedprocess

[30–32].The labeledantibody should surpass thebound fraction,

butit isboundatthesurface Theeffect becameeven more

pro-nouncedwhentheunlabeledantibodywasloadedagainontopof

thelabeledantibody.Thegreenfractionmigratedfromthesurface

area towardsthecoreofthebeads(Fig.8,avideoofthis

experi-mentincludingtheloadingphaseandelutionphaseisavailablein

the onlinesupplementary material) We assume that the

adsorp-tionisnot fullybasedona porediffusion-limitedprocess;itmay

be a combinationof solidandporediffusion.Therefore,a certain

portionoflabeled tracerremains onthesurfaceandconstant

ex-changebetweenliquidandsolidphasesoccurs

Fig 8 Resin beads under the microscope in a microcolumn A: Fresh resin Beads

are shown in black and white B: Unlabeled antibody was loaded (which is not vis- ible) C and D: Labeled antibody was loaded (green) E: Unlabeled antibody was loaded (not visible) Please see the video of this process online in the supplemen- tary material

Fig 9 Pulse injection in batch chromatography using IgG1 The blue line shows the unlabeled antibody concentration The green line shows the labeled antibody concen-

tration

Tovalidatepreviousresults,thelastsetofexperimentswas

per-formed usingan antibodyofsubclass1andthesamefluorescent

dye.Intwoseparateexperiments,onepulseoftracerwasinjected

at the start and the end of the loading phase The experiments

wereperformedinbatchmodeasdescribedin2.5

Whenthelabeledmaterialwasinjectedatthestartofthe

load-ing phase, only91% werefound in theelution peakeven though

thehypothesiswastofind100%ofinjectedmaterialintheelution

peak When the labeled material was injected at the end ofthe

loadingphase,62%werefoundintheelutionpeakeventhoughthe

hypothesiswastofind35%ofinjectedmaterialintheelutionpeak

(asitwasinjectedat65%ofbreakthrough)(Fig.9).Theresults

val-idatethepreviousexperimentswithIgG2,thoughtheconstant

ex-change occurred to a lesserextent The reason is that the

bind-ing affinity of IgG1towardprotein Aresin is higherthan that of

IgG2, solessexchangewasobservedbetweentheliquidandsolid

phases.However,theexchangeisstillsignificant

The dyebinding siteon theantibodyisimportantandcan

in-dicateifitisaffectingtheproteinAbinding.Inourcase,weused

mab SelectSure, which is composed of four mutatedZ-domains

[33].It is a myththat the Z-domain derived fromstaphylococcal

protein Aexclusivelybinds totheFc-domainofantibodybecause

theelution pHisindependentoftheantibodysubclass[34].Bach

et al.investigatedthebinding behavior of15 differentantibodies

of subclass IgG1and IgG2 andfound a secondary interaction by

incubation of the F(ab’)2 fragments for all of them The weaker

secondary interaction also suggestsa constant exchange between

bound andfreeantibodies[35].Thisobservationcorroboratesour

findings

4 Conclusion

We conclude that a fluorescent-labeledantibody serves as an

inert tracer in protein A affinity chromatography for

determin-ing the RTD only when it is performed at a high salt

concen-tration When we usethe fluorescent-labeledantibodyunderthe

mentioned conditions,anysection of theinlet distributesequally

through the elution peak, meaning that the ratio of labeled and

unlabeled antibodies remains the same throughout the elution

peak Although the constant exchange of antibodiesbetweenthe

liquid andsolid phase in the loadingphase complicates the RTD

determination, the adsorption process is a combination of solid

and porediffusion,which explainsthe exchange ofmaterial

dur-ing loading Thisalso suggeststhat, in thiscase, we have an

ex-tremelywideRTD.Inthecaseofthecirculationofproductin

con-tinuouschromatography unit operation, such asPCC,the circula-tionofproductinthesystemincreasestheaverageresidencetime

Narges Lali: Execution of experiments, design of the exper-iments, drafting of the manuscript, interpretation of the data

Peter Satzer:designoftheexperiments,interpretationofthedata:

Alois Jungbauer:Conceptualization,Resources,Writing Review& Editing,Supervision,Fundingacquisition

Declaration of Competing Interest

Theauthorsdeclarethattheyhavenoknowncompeting finan-cialinterestsorpersonalrelationshipsthatcouldhaveappearedto influencetheworkreportedinthispaper

Acknowledgments

This work was supported by the European Union’s Horizon

2020research andinnovationprogramundergrantagreementno 635557

Supplementary material associated with this article can be found,intheonlineversion,atdoi:10.1016/j.chroma.2022.463530

References

[1] O Levenspiel, Chemical Reaction Engineering, 3rd ed., Wiley, New York, 1999 ed[Online] Available: http://app.knovel.com/web/toc.v/cid:kpCREE0 0 05 [2] W Engisch, F Muzzio, Using residence time distributions (rtds) to address the traceability of raw materials in continuous pharmaceutical manufacturing, J Pharm Innov 11 (2016) 64–81, doi: 10.1007/s12247-015- 9238- 1

[3] N Lali, A Jungbauer, P Satzer, Traceability of products and guide for batch def- inition in integrated continuous biomanufacturing, J Chem Technol Biotech- nol (2021), doi: 10.1002/jctb.6953

[4] "GUIDANCE DOCUMENT quality considerations for continuous manufac- turing issued by: office of medical products and tobacco, center for drug evaluation and research document Nr FDA-2019-d-0298 February 2019." https://www.fda.gov/regulatory-information/search-fda-guidance-documents/ quality- considerations- continuous- manufacturing (accessed

[5] J Sencar, N Hammerschmidt, A Jungbauer, Modeling the residence time dis- tribution of integrated continuous bioprocesses, Biotechnol J 15 (8) (2020) e20 0 0 0 08 Aug, doi: 10.10 02/biot.2020 0 0 0 08

[6] S Mun, Y Xie, N.-H.L Wang, Residence time distribution in a size-exclusion SMB for insulin purification, AlChE J 49 (8) (2003) 2039–2058, doi: 10.1002/ aic.690490814

[7] S Pawlowski, N Nayak, M Meireles, C.A.M Portugal, S Velizarov, J.G Crespo, CFD modelling of flow patterns, tortuosity and residence time distribution in monolithic porous columns reconstructed from X-ray tomography data, Chem Eng J 350 (2018) 757–766 2018/10/15/, doi: 10.1016/j.cej.2018.06.017

N Lali, P Satzer and A Jungbauer Journal of Chromatography A 1683 (2022) 463530

[8] M.A Teeters, I Quiñones-García, Evaluating and monitoring the packing behav-

ior of process-scale chromatography columns, J Chromatogr A 1069 (1) (2005)

53–64 2005/03/25/, doi: 10.1016/j.chroma.2005.02.051

[9] Y.N Sun, et al., Comparison of protein A affinity resins for twin-column contin-

uous capture processes: process performance and resin characteristics, J Chro-

matogr A 1654 (2021) 462454 Sep 27, doi: 10.1016/j.chroma.2021.462454

[10] C.-.S Chen, N Yoshimoto, S Yamamoto, Prediction of the performance of

capture chromatography processes of proteins and its application to the re-

peated cyclic operation optimization, J Chem Eng Jpn 53 (11) (2020) 689–

697, doi: 10.1252/jcej.20we116

[11] W Krepper, P Satzer, B.M Beyer, A Jungbauer, Temperature dependence of an-

tibody adsorption in protein A affinity chromatography, J Chromatogr A 1551

(2018,) 59–68 May 25, doi: 10.1016/j.chroma.2018.03.059

[12] E.X Perez-Almodovar, G Carta, IgG adsorption on a new protein A adsorbent

based on macroporous hydrophilic polymers I Adsorption equilibrium and

kinetics, J Chromatogr A 1216 (47) (2009) 8339–8347 Nov 20, doi: 10.1016/j

chroma.2009.09.017

[13] J.T McCue, G Kemp, D Low, I Quiñones-Garcı ´a, Evaluation of protein-A chro-

matography media, J Chromatogr A 989 (1) (2003) 139–153, doi: 10.1016/

s0 021-9673(03)0 0 0 05-0

[14] R Hahn, P Bauerhansl, K Shimahara, C Wizniewski, A Tscheliessnig, A Jung-

bauer, Comparison of protein A affinity sorbents: II Mass transfer properties, J

Chromatogr A 1093 (1) (2005) 98–110 20 05/11/04/, doi: 10.1016/j.chroma.20 05

07.050

[15] J Weinberg, S Zhang, G Crews, E Healy, G Carta, T Przybycien, Polyclonal

and monoclonal IgG binding on protein A resins-evidence of competitive bind-

ing effects, Biotechnol Bioeng 114 (8) (2017) 1803–1812 Aug, doi: 10.1002/bit

26286

[16] H.S Fogler, Elements of Chemical Reaction Engineering (in English), 2006

[17] A Jungbauer, Continuous downstream processing of biopharmaceuticals,

Trends Biotechnol 31 (8) (2013) 479–492 Aug, doi: 10.1016/j.tibtech.2013.05

011

[18] A.L Cataldo, D Burgstaller, G Hribar, A Jungbauer, P Satzer, Economics and

ecology: modelling of continuous primary recovery and capture scenarios for

recombinant antibody production, J Biotechnol 308 (2020) 87–95 Jan 20,

doi: 10.1016/j.jbiotec.2019.12.001

[19] C.A Teske, R Simon, A Niebisch, J Hubbuch, Changes in retention behavior of

fluorescently labeled proteins during ion-exchange chromatography caused by

different protein surface labeling positions, Biotechnol Bioeng 98 (1) (2007)

193–200 Sep 1, doi: 10.1002/bit.21374

[20] J Gomis-Fons, N Andersson, B Nilsson, Optimization study on periodic

counter-current chromatography integrated in a monoclonal antibody down-

stream process, J Chromatogr A 1621 (2020) 461055 Jun 21, doi: 10.1016/j

chroma.2020.461055

[21] C.A Teske, M Schroeder, R Simon, J Hubbuch, Protein-labeling effects in con-

focal laser scanning microscopy, J Phys Chem B 109 (28) (2005) 13811–13817

Jul 21 + , doi: 10.1021/jp050713

[22] C.A Teske, E von Lieres, M Schroder, A Ladiwala, S.M Cramer, J.J Hubbuch,

Competitive adsorption of labeled and native protein in confocal laser scan-

ning microscopy, Biotechnol Bioeng 95 (1) (2006) 58–66 Sep 5, doi: 10.1002/

bit.20940

[23] U Resch-Genger, M Grabolle, S Cavaliere-Jaricot, R Nitschke, T Nann, Quan- tum dots versus organic dyes as fluorescent labels, Nat Methods 5 (9) (2008) 763–775 Sep, doi: 10.1038/nmeth.1248

[24] S Vira, E Mekhedov, G Humphrey, P.S Blank, Fluorescent-labeled antibodies: balancing functionality and degree of labeling, Anal Biochem 402 (2) (2010) 146–150 Jul 15, doi: 10.1016/j.ab.2010.03.036

[25] R Godawat, K Brower, S Jain, K Konstantinov, F Riske, V Warikoo, Periodic counter-current chromatography – design and operational considerations for integrated and continuous purification of proteins, Biotechnol J 7 (12) (2012) 1496–1508 Dec, doi: 10.10 02/biot.20120 0 068

[26] M Salvalaglio, L Zamolo, V Busini, D Moscatelli, C Cavallotti, Molecular mod- eling of protein A affinity chromatography, J Chromatogr A 1216 (50) (2009) 8678–8686 2009/12/11/, doi: 10.1016/j.chroma.2009.04.035

[27] W.L DeLano, M.H Ultsch, A.M de Vos, J.A Wells, Convergent solutions to binding at a protein-protein interface, Science 287 (5456) (20 0 0) 1279–1283, doi: 10.1126/science.287.5456.1279

[28] L Yang, J.D Harding, A.V Ivanov, N Ramasubramanyan, D.D Dong, Effect of cleaning agents and additives on Protein A ligand degradation and chromatog- raphy performance, J Chromatogr A 1385 (2015) 63–68 Mar 13, doi: 10.1016/j chroma.2015.01.068

[29] J Hubbuch, T Linden, E Knieps, J Thömmes, M.-.R Kula, Mechanism and ki- netics of protein transport in chromatographic media studied by confocal laser scanning microscopy: part II Impact on chromatographic separations, J Chro- matogr A 1021 (1) (2003) 105–115 20 03/12/22/, doi: 10.1016/j.chroma.20 03.08

092 [30] M Zhu, G Carta, Protein adsorption equilibrium and kinetics in multimodal cation exchange resins, Adsorption 22 (2) (2016) 165–179 2016/02/01, doi: 10 1007/s10450-015-9735-z

[31] E.X Perez-Almodovar, Y Wu, G Carta, Multicomponent adsorption of mono- clonal antibodies on macroporous and polymer grafted cation exchangers, J Chromatogr A 1264 (2012) 48–56 2012/11/16/, doi: 10.1016/j.chroma.2012.09

064 [32] T Liu, J.M Angelo, D.-.Q Lin, A.M Lenhoff, S.-.J Yao, Characterization

of dextran-grafted hydrophobic charge-induction resins: structural proper- ties, protein adsorption and transport, J Chromatogr A 1517 (2017) 44–53 2017/09/29/, doi: 10.1016/j.chroma.2017.07.090

[33] M Linhult, et al., Improving the tolerance of a protein a analogue to re- peated alkaline exposures using a bypass mutagenesis approach, Proteins Struct Funct Bioinf 55 (2) (2004) 407–416 2004/05/01doi: 10.1002/prot.10616, doi: 10.1002/prot.10616

[34] S Ghose, M Allen, B Hubbard, C Brooks, S.M Cramer, Antibody vari- able region interactions with Protein A: implications for the development

of generic purification processes, Biotechnol Bioeng 92 (6) (2005) 665–673 2005/12/2010.1002/bit.20729, doi: 10.1002/bit.20729

[35] J Bach, N Lewis, K Maggiora, A.J Gillespie, L Connell-Crowley, Differential binding of heavy chain variable domain 3 antigen binding fragments to pro- tein a chromatography resins, J Chromatogr A 1409 (2015) 60–69 2015/08/28/, doi: 10.1016/j.chroma.2015.06.064