Mechanical characterisation of agarose-based chromatography resins for biopharmaceutical manufacture

Mechanical characterisation of agarose-based resins is an important factor in ensuring robust chromatographic performance in the manufacture of biopharmaceuticals. Pressure-flow profiles are most commonly used to characterise these properties.

jou rn al h om ep a g e : w w w e l s e v i e r c o m / l o c a t e / c h r o m a

Mauryn C Nwekea, R Graham McCartneyb, Daniel G Bracewella,∗

a r t i c l e i n f o

Keywords:

Pressure-flow

a b s t r a c t

Mechanicalcharacterisationofagarose-basedresinsisanimportantfactorinensuringrobust chro-matographicperformanceinthemanufactureofbiopharmaceuticals.Pressure-flowprofilesaremost commonlyusedtocharacterisetheseproperties.Thereareanumberofdrawbackswiththismethod, includingthepotentialneedforseveralre-packstoachievethedesiredpackingquality,theimpactof walleffectsonexperimentalsetupandthequantitiesofchromatographymediaandbuffersrequired

Toaddresstheseissues,wehavedevelopedadynamicmechanicalanalysis(DMA)techniquethat char-acterisesthemechanicalpropertiesofresinsbasedontheviscoelasticityofa1mlsampleofslurry Thistechniquewasconductedonsevenresinswithvaryingdegreesofmechanicalrobustnessandthe resultswerecomparedtopressure-flowtestresultsonthesameresins.Resultsshowastrong correla-tionbetweenthetwotechniques.Themostmechanicallyrobustresin(CaptoQ)hadacriticalvelocity 3.3timeshigherthantheweakest(SepharoseCL-4B),whilsttheDMAtechniqueshowedCaptoQto haveaslurrydeformationrate8.3timeslowerthanSepharoseCL-4B.Toascertainwhetherpolymer structureisindicativeofmechanicalstrength,scanningelectronmicroscopyimageswerealsousedto studythestructuralpropertiesofeachresin.ResultsindicatethatDMAcanbeusedasasmallvolume, complementarytechniqueforthemechanicalcharacterisationofchromatographymedia

©2017TheAuthor(s).PublishedbyElsevierB.V.ThisisanopenaccessarticleundertheCCBYlicense

(http://creativecommons.org/licenses/by/4.0/)

1 Introduction

Manufacturersmustensurethatchromatographymediameet

a broad range of requirements before use for the

sepa-ration/purification of biological products These requirements

include a number of safety considerations (leachables,

toxicol-ogy),performance(capacity,specificity,throughput),cost(capital

investment,longevity)andstability,amongothers[1].Stabilitycan

besplitbroadlyintotwocategories−chemicalandmechanical.The

chemicalresistanceofchromatographymediaisdependentonthe

couplingchemistryaswellasthechoiceofspacerandligand

chem-istryandstability.Whereas,themechanicalstabilityisdependent

largelyonthechoiceandcompositionofthebasematerial,particle

sizedistribution,particleporosity,andtoalesserextent,ligandand

liganddeployment[2,3]

Thebasematerialischosenbasedonanumberoffactorssuch

ascost,thepropertiesofthematerialtobeprocessedandsurface

areaandmasstransfercharacteristics,givingrisetoparameters

such as dynamic binding capacity (DBC) maximum flow rates, maximumnumberofcyclesetc.Basedonthis,different manufac-turersusedifferentcompositematerialsfortheirchromatographic media[4].Agaroseis acommonlyusedbasematrixmaterialin biopharmaceuticalpurificationasitrelativelystraightforwardto manufacture and customise certainproperties suchas porosity andspecificbindingproperties.Thispaperfocusesparticularlyon MabSelectTM,SepharoseTM and CaptoTM media(GEHealthcare, Uppsala,Sweden)

Agaroseisoneoftwomainconstituentsofagarandisgenerally extractedfromseaweed.Itiscomposedofapolysaccharide poly-mermaterialformedofrepeatingunitsof1–3-linked␤-Dgalactose and1,4-linked3,6-anhydro-␣-l-galactose[5].Oncetheagarhas beenprocessed,theagaroseisintheformofadrypowder.Itis thendissolvedinanaqueoussolution>85◦C,causingthechains

todegrade[3,6].Whenthesolutionreachesacertainviscosity,it

iscooledand poured,whilstsimultaneouslybeingstirredintoa non-polarorganicsolventwhichcontainsanemulsifier.These con-ditionsinducetheformationofsphericalbeads(emulsification) Thestirringandcoolingratesareakeyparametersindetermining certainstructuralcharacteristicssuchasporosity,poresize

distri-https://doi.org/10.1016/j.chroma.2017.11.038

Fig 1. General method for making porous agarose beads The agarose solid is dissolved in water heated to about 90 ◦ C This is then added to a stirred vessel containing a

butionandparticlesizedistribution,whichtendstorangefrom20

to300␮m[4](Fig.1)

Uponformation,thebeadsareinsolubleandsedimentintothe

higher density water phase, as opposed tothe organicsolvent

phase Thebeads are subsequently cross-linked witha reagent

suchasepichlorohydrin.Theextenttowhichthisisdoneisone

ofthecritical factors thatdeterminetherigidity of thematrix

However,cautionmustbetakenatthisstepasover-cross-linking

mayreduceporosity,liganddeploymentandcompressibility

char-acteristics[7,8].Whentheprocessiscompleted,theresincanbe

usedinvariousapplicationssuchassizeexclusionanddesalting.It

mayalternativelygoontobefunctionalisedwithdifferentligand

chemistries,afterwhichitcanbeusedinanumberofmodesfor

variousbiopharmaceuticalapplications[1,3,9–11]

Toensureconsistencyinthestructuralandmechanical

prop-erties of chromatography media, the media has to be well

characterised.Structuralintegritytestinginvolveslookingatpore

sizeandparticlesizedistributionsandporosity,whichcan

gener-allybeascertainedindirectlybyobservingtitrationcurvesorstatic

capacity.Therehavealsobeenreportsondevelopinglab-based

pro-ceduresthatinvolvetheuseofmicromanipulation[9,12].Abetter

ideaofmechanicalandcolumnperformanceisusuallydetermined

bypressure-flowcharacterisation.Thistechniqueinvolves

gradu-allyincreasingtheflowrateandobservingariseinthepressure

profileinthecolumn.Atacertainflowvelocity,thepressureinthe

columnwillcontinuetorisewithoutfurtherincreasetotheflow

rate.Itisatthispointthatthecriticalvelocityhasbeenreachedand

thecolumnhas‘failed’[13]

Theadvantagesofthismethodincludetheabilitytodetermine

thebehaviourofchromatographymediainapackedbedandhow

mechanicalpropertiesvarywithmediaviscosity,pH,ionicstrength

etc.However,adrawbackofthismethodisthatitrequiresthatthe

operatoradherestostringentpackingcriteriatoobtainmeaningful

data.Whenpackingcolumns,severalre-packsmayberequiredto

achievethedesiredasymmetryandeachresin,dependingonits

chemicalandmechanicalproperties,hasitsownspecificpacking

criteria.Furthermore,itisnecessarytouseacolumnofasuitable

diameter,suchthatwalleffectsthatsupporttheresininnarrow

columnsdonotdominate[14].Thebedheightalsoneedstobe

rep-resentative,aspressuredropdirectlycorrelatestotheheightofthe

bed,meaningheightsof15cmorgreateraretypicallyused[15].For

thesereasons,thepressure-flowtechniqueconsumeslarge

quan-titiesofchromatographicmediaandbuffers,whichiscostly[16]

To address these drawbacks, we have developed the use

of dynamic mechanical analysis (DMA) (Fig.2).This technique

involvesapplyingasmalldeformationtoasampleinacyclic man-nerandallowsforthesamplematerialtorespondtochangesin stress,temperature,strain,frequency,forceaswellasother param-eters.Itisusedwidelyinthebioengineeringsectorandthefieldof biosciencestocharacterisetheviscoelasticpropertiesofvarious biologicaltissueand otherbiomaterials.Traditionally,thestress andstrainparametersareusedtocalculateYoung’smodulusto giveanindicationofchangesinelasticproperties.Moronietal.,

2006[17]usedthetechniquetoinvestigatetheuseofscaffoldsto mimichumantissue.Theyfoundthatthetechniquewas partic-ularlysensitivetoporesizechangesinscaffolds.Withincreasing porosityinthescaffolds,therewasadecreaseinelasticproperties, whichcorrespondedtoanincreaseinstrain.Ithasalsobeenused

tolookatthemechanicalpropertiesofmaterialssimilartoagarose gels,suchashydrogels.Meyvis&Stubbe2002[18]usedDMAasa comparativetechniquetoshearrheometrytoinvestigate mechan-icalpropertiesofpharmaceuticalhydrogels.Theyfoundastrong correlationbetweenthetwotechniquesbutobservedthatDMA canbeusedtoinvestigatemanymoremechanicalparametersthan solelyviscoelasticity

We have applied the use of DMA to investigate the vis-coelastic properties of small quantities of seven agarose-based

Table 1

Sepharose 4 FF, Sepharose 6 FF, Q Sepharose HP 400 cm/hr − 2 min

resins,namely:SepharoseCL–4B(SCL4B),Sepharose4FF(S4FF),

SepharoseCL–6B(SCL6B),Sepharose6FF(S6FF),Q-SepharoseHigh

Performance (Q-HP), MabSelectTM (MabSelect) and CaptoTM Q

(CaptoQ)(GEHealthcare,Uppsala,Sweden).Weinvestigatehow

the slurriesrespond to strain over a fixed period of time We

thenlooktodrawcorrelationsbetweentheresultsobtainedfrom

pressure-flowandDMAexperimentstoascertainwhetherDMA

canbeusedascomplementarytechniqueforthemechanical

char-acterisationofchromatographymedia

2 Materials and methods

2.1 Pressure-flow

2.1.1 Equipment

Abench-scalecolumnwithadjustablecolumnlengthandinner

diameterof1.6cm(modelXK16,GEHealthcare,Uppsala,Sweden)

wasused.ThiswasoperatedontheAKTAPure(GEHealthcare,

Upp-sala,Sweden).Columnpressuredrop(P)wasmeasuredusing

theinternalpressuremeasurement devicesinstalledinthefeed

deliverysystemoftheAKTAPureandthevolumetricflowratewas

measuredmanuallyusingthemethodemployedby[15]

2.1.2 Chromatographymedia

Sepharose CL-4B, Sepharose CL-6B, Sepharose 4 Fast Flow,

Sepharose6 FastFlow,Q SepharoseHighPerformance,

MabSe-lectandCaptoQ(GEHealthcare,Uppsala,Sweden)wereusedin

thisstudy.Theseagarose-basedchromatography resinshavean

averageparticlesizeof 80␮m, withabeadsize distributionof

between24and165␮m.Themechanicaldifferencesbetweenthe

sevenresinslieintheagarosecontentandtheextentofstructural

cross-linkingpresent

MabSelectandCaptoQaremadeofhighlycross-linkedagarose,

whereas Sepharose CL-4B/CL-6B, Sepharose 4 FF/6 FF and Q

Sepharose HP are structurally simplerin terms of their

cross-linking However Sepharose 6 FF, CL-6B, Q Sepharose HP and

MabSelectallcontainthesamepercentageofagaroseintheir

matri-ces(6%),whileSepharoseCL–4BandSepharose4FFcontain4%and

CaptoQ7%

2.1.3 Procedure

Packing–Allchromatographymediawasmadeupto50%slurry

concentration The same procedurewas repeated for all seven

resins.30mlofslurrywaspouredintothecolumnandallowedto

gravitysettleovernight.Theadaptorwasloweredintothe

super-natanttostarttheflowpack.Allcolumnswerepackedat15cm/hr

for60minandsubsequentlyat30cm/hrfor30min.Thecolumns

showninTable1werefurtherconsolidated.Thetopadaptorwas

thenloweredtothetopofthebed.Thepackingmediumusedfor

allbufferswasdistilledH2O(dH2O)

Performancetesting – 2% v/vof acetonewas measuredand

addedinto30mlofdH2Oina50mlfalcontube(CELLSTAR®,UK)

1ml of this solution wasinjected into a 600␮l loop and then

loadedontothecolumn.TheeluentusedinthisstudywasdH2O

at30cm/hr.Apeakwasthengeneratedwithin30min.The

asym-metrywascalculatedusingthein-builtfunctionontheUnicorn6.4

software

Pressure-flowmethod–Theflowrateofthepackingbufferwas continuallygraduallyincreaseduntila35kPaincreaseinpressure dropwasobserved,asdescribedbyTranetal.,2007[15].Atthis point,theflowrateand anychangesinbedheightwere manu-allyrecorded.Atacertainflowrate,thepressurebegantoincrease exponentiallywithnofurtherchangetotheflowvelocity.Atthis pointitwasdeemedthatthecriticalvelocityforthecolumnhad beenreached

2.2 Dynamicmechanicalanalysis 2.2.1 Column/holderdesign

10identicalcylindricalblocksoftransparentacrylicweredrilled withaninnerdiameterof11mm,anouterdiameterof14mmand

a heightof 15mm.Thebottomwaswrappedina thinsheetof parafilm(0.1mmthickness)tocontaintheslurry

2.2.2 Samplepreparation

Analiquot of 10ml ofeach resin wasplaced intoa labelled

50mlfalcontubeandcentrifugedfor5minat3000rpm (Eppen-dorfcentrifuge5810R,ThermoFisherScientific,UK)andtheslurry concentrationwasnotedbasedonthevolumeratioofliquid to slurryin thefalcon tube.Thestorage buffer(20%ethanol) was decanted,replacedwiththeirrespectivepackingbuffersandthe slurrysolutionwasmadeuptoa70%slurryconcentration.The aliquotswereresuspendedandtheprocedurewasrepeateduntil thestoragebufferhadbeencompletelyremoved.1.42mlofeach aliquotwaspipettedintotheirrespectivelylabelledholderandleft

tosettleovernight,suchthattheirsettledbedheightwas1cm.A consistentslurryconcentrationisimportantinachievingauniform settledbedvolumeandheightforcomparablestrainmeasurements acrossallresins

2.2.3 DMAprocedure DMAwascarriedoutontheDMA7ehardware,withaTAC7/DX controllerandPyrisManagersoftware(PerkinElmer,UK).Inthis proceduretheforcereadingiszeroed,theweightoftheprobeis taredandtheprobepositioniszeroedwhenthelidisloweredtothe baseofthepan.Thelidisliftedandtheholdercontainingtheslurry

isplacedontothepan.Thelidisloweredtothetopoftheresinbed, theheightisreadandthemethodologyisstarted.Inthis method-ology,thelidappliesaforceof100mN/minatafrequencyof1Hz for80minandatime-strainplotisgeneratedsimultaneously.Upon completionofthemethodology,theslopeofthelineismanually fit-tedfromtheorigintothepointbeforeultimatecompression(Fig.6) usingthein-builtslopefunctioninthePyrisManagersoftware 2.3 Scanningelectronmicroscopy

Allsampleswerecriticalpointdriedandimagedusingthesame protocoldescribedinNwekeetal.,2016[19]

3 Results & discussion

Theresinsusedinthisstudywereselectedbasedontheir dif-ferencesinpercentageofagarosecontentanddifferencesobserved

intheirfibrousstructure,poresizedistributionandcross-linking viascanningelectronmicroscopy(Fig.5,Table2).Their mechani-calpropertiesarecharacterisedusingthestandardpressure-flow methodandthiswillthenbecomparedtoresultsfromdynamic mechanicalanalysis

3.1 Pressure-flow

Inthis technique,theflowrateismanuallyincreaseduntila runawayriseinthepressureprofileisobserved.Theflowrateat

Table 2

Resin Bead size range (␮m) Pore size range (nm) % agarose Extent of cross-linking

whichthisoccursisconvertedtolinearvelocity.Thisisthepoint

atwhichthecolumnhas‘failed’andistermedthecriticalvelocity

Themorerigidtheresinis,thehigherthecriticalvelocity.Three

repeatsoftheprocedure(section2.1)wereconductedforallseven

resinsandtheaveragesareplotted.Thestandarddeviationsfrom

theaverage(basedontherepeats)arerepresentedbytheerror

bars(Fig.3).Thedegreetowhichanyindividualrepeatmayvary

isreliantmainlyoncolumnpackingandtheresultingasymmetry

Theprobabilitythatacolumnwillpackinexactlythesameway,

despiteusingthesameprocedureislow.Thisisrepresentedbythe asymmetryvalueobtained.Althoughtheasymmetrydifferedforall repeats,itwasmaintainedintherangeof0.8-1.2(whichmayhave requiredmultiplerepackstoachieve).Areductioninbedheight duringthechangestoflowvelocitymayalsobeobserved[15] Fig 4 shows the critical velocities for each resin using the pressure-flowcharacterisation techniqueusinganXK16column withabedheightof15cm.CaptoQhasthehighestcritical veloc-ityat492cm/hr,followedbyMabSelect−423cm/hr,Q-Sepharose

HP−353cm/hr,Sepharose6FF−348cm/hr,SepharoseCL–6B−

283cm/hr, Sepharose4FF −204cm/hr and SepharoseCL–4B−

149cm/hr

TheresultsshowthatCaptoQistheleastcompressibleofthe7 resins,followedbyMabSelect.Thisisexpectedasbothresinsare madeofhighlycross-linkedagarosepolymersandcontain7%and 6%agaroserespectively.Q-HPandS6FFarecross-linkedresinsthat contain6%agarose.Thesetworesinshavequasi-identicalcritical velocities.Theirmainstructuraldifferencesareobservedintheir averagebeadsizeandtheiraverageporesizedistributions(Fig.5, Table2)sotheirdynamicbehaviourinthecolumnisnotexactlythe same(Fig.3 –Q-HPexhibitsslightlyhigherpressuredrop).S6FF resinsare2–3timeslargerinsize(dp)comparedtoQ-HP, how-evertheaverageporesizeofS6FF(andFastFlowresinsingeneral)

isapproximately2–3timessmallerthanthatofQ-HP[20–22].It hasbeenestablishedthatbothporesizedistributionandbeadsize contributetothemechanicalpropertiesofchromatographymedia [23,9].Thetrade-offbetweenthesetwoparameters,aswellastheir identicalmechanicaltraits,mayexplainwhybothresinshavevery similarcriticalvelocityvalues

Theresultsalsoshowthedifferencesin mechanicalstrength betweenSCL6BandS6FF,aswellasSCL4BandS4FF.Bothpairsof resinsarecross-linkedandcontain6%and4%agaroseintheir matri-cesrespectively,howeverbothfastflowresinsaremechanically

Fig 5. Scanning electron micrographs showing: Sepharose 4FF (4% cross-linked agarose), Sepharose CL–4 B (4% cross-linked agarose), Sepharose 6FF (6% cross-linked agarose),

strongerthantheir–CLcounterparts.Inbothcases,thefastflow

resinswithstandmuchhigherflowratesaccordingtothe

manu-facturer’sspecification,whichmayindicatethattheircross-linking

wasmoreextensive.Scanningelectronmicrographswereobtained

toshowthestructuralpropertiesofeachresin,whichwereusedto ascertainwhetherpolymerstructure isindicativeofmechanical strength.Theirmicrographsshowthattheyarestructurally differ-ent(Fig.5).SCL6B,forexample,appearstobemorefibrousand

Fig 6.Schematic of DMA methodology The lid is equipped with a sensor that records the initial height of the sample When the methodology is started the descending lid

morediscontinuouscomparedtoitsfastflowcounterpartwhich

hasamorehomogenous,continuousstructure,whichmay

indi-categreatermechanicalstrength.Itshouldbenotedthatthecited

literatureplots[14],[15]depictaxeswithvaryingmetrics.These

citationsareusedtodescribethemethodbywhichpressure-flow

characterisationwascarriedoutinthisstudy.GEHealthcaredata

sheetsdepictsimilarprofilestotheonesobtainedin thisstudy

[25],[26].Differencesincritical velocityvaluesreportedin this

studymaybe attributedtodifferent packingtechniquesin the

differentcolumnsused[27].Thepressure-flowprofilesobtained

byGEHealthcareuseproductionscalecolumns(AxiChrom,BPG)

andmainlypack-in-placeandaxialmechanicalcompression

pack-ing,whilstthisstudyuseslab-scaleXK16columnsundertheflow

packingtechnique.Itisworthhighlightingthatthemostcritical

aspectofthisstudyisthatthesamebatchofresinswereusedin

theapplicationofalltechniquesreportedinthisstudy

3.2 Dynamicmechanicalanalysis(DMA)

Thistechniquecharacterisesmechanical propertiesbasedon

theviscoelasticityofasmallsampleofresin.ConventionalDMAis

usedtocharacterisehomogeneouslyshapedbiomaterialsto

deter-mineelasticpropertiessuchasYoung’smodulus[17]however,we

haveadaptedthetechniquesuchthatitcharacterisesthe

proper-tiesofaslurry.Theequipmentiscomposedmainlyofapananda

lid,equippedwithsensors.Thelidinparticularisequippedwith

a sensor thatallows it tostopjust atthesurface of theslurry

onceindescent.Whenthemethodologyisstarted,a sinusoidal

forceisappliedataconstantfrequency.Astheliddescends,the

slurrymovesaround thesidesofthe lidandthis movement is

recordedasadisplacementpercentagewithtime Meanwhile,a

time-strainprofileisgenerated,wherestrainisthedisplacement

ofthelidthroughtheresinbedrecordedasapercentage.Oncea

strainthresholdisexceeded,therateofincreaseisvastlyreduced

ortheplotbeginstoleveloutcompletelyandatthispointultimate

compressionisreached.Thisiseitherwhenthelidhashitthe

bot-tomofthepan,orwhenlittleornofurtherdeformationoftheslurry

canbeachievedwithconstantforce.Forconsistency,theslopeof

thelineistakenbeforeultimatecompressionandthisprovides

informationaboutthemovementofthelidthroughtheslurrywith

constantforce.AstrainversustimeplotisgeneratedbythePyris Managersoftwareandtheslopeofthelinebeforeultimate com-pressiondeterminestheslurrydeformationrate.Theslopeofthe lineismanuallyfittedfromtheorigintothepointbeforeultimate compressionusingthein-builtslopefunctioninthePyrisManager software.Theunitsarerecordedas%/min(Fig.6).Thelessviscous themediais,thequickerthelidwillmovethroughtheslurry, there-forethehigherthe%strainperminute.Theprocedureisrepeated

3timesforeachresin(Fig.7)

Fig.8ashowstheslurrydeformationratesforeachresinusing theDMAtechnique.CaptoQhastheslowestslurrydeformation rateat0.36%/min,followedbyMabSelect–0.55%/min,Q-Sepharose

HPandSepharose6FF–1.1%/min,SepharoseCL–6B–1.7%/min, Sepharose4FF–2.5%/minandSepharoseCL–4B–3%/min Fig.8 showsthegraphobtainedbyplottingthereciprocalof theSDRvaluesforeachresin.Theresultingparameterwastermed

‘slurryresistance(1/%min−1)’.Thiswasdonetomoreclearlyshow the trend between the pressure-flow technique and the DMA technique with particular focus on the gapbetween the more mechanicallyrobustresinsandtheweakerresins.Theresultsshow CaptoQhasthehighestslurryresistanceof2.8,followedby Mab-Select−1.81,Q-HPandS6FF−0.90,SCL6B−0.59,S4FF−0.4and SCL4B−0.3

ThesametrendsobservedinFig.3areobservedinFig.8b.The resultsfromFig.8 showthatCaptoQismostresistantto deforma-tion,followedbyMabSelect,Q-SepharoseHighPerformanceand Sepharose6FastFlow,SepharoseCL-6B,Sepharose4FastFlow and SepharoseCL-4B Similartothetrendsobserved insection 3.1,theresultsalsoshowthatQ-SepharoseHighPerformanceand Sepharose6FastFlowexhibitverysimilarviscoelasticproperties Bothresinscontain6%agarosehowevertherearedifferencesin theiraveragebeadsizesandporesizes.Poresizeisinfluencedbya numberoffactors,includingtheextentofcrosslinkingwhich influ-encesmechanicalrigidity[1].Thisbecomesofrelevancewhenthe beadsmovepasteachotherthroughthegapsastheliddescends andthegapsizeof500␮mislargeenoughsuchthatitdoesnot allowforradialrestriction/compressionofsinglebeads.The trade-offbetweenthefactthatQ-HPhasalargeraverageporesizethan S6FF,butS6FFhasalargerbeadsizerange,wouldmeanthatthere arefewerS6FFbeadsforthesamegivenvolume.Thiscouldexplain

Fig 7. (a) Strain v time plot for 3 repeats of Sepharose CL–4 B (4% cross-linked agarose) (one solid line, one dashed line, one dotted line) (b) Averages of 3 out of 7 resins –

points.

thedifferencesinviscoelasticpropertiesbetweenSCL6BandS6FF,

aswellasSCL4BandS4FF.Furtherexplanationsfortheobserved

trendsareoutlinedinSection3.1.ThedatafromFigs.4and8are

plottedtoestablishatrend(Fig.9)

3.3 Datacorrelation–pressure-flowvsdynamicmechanical

analysis

Fig.9a shows a strongnegative correlationbetween critical

velocityandslurrydeformationratefor all7resinsusedinthis

study.ThismeansthatthestrongerresinssuchasMabSelectand

CaptoQhavelowSDRsandhighucritvaluesandtheoppositeistrue

formechanicallyweakerresinssuchasSCL4BandS4FF.Thetrend

beginsinalinearfashionwiththefirstfiveresinsbutthentailsoff

whenthemoremechanicallyrobustresinsappear.This

represen-tationofresultsindicatesthatastheresinsbecomemechanically

moresimilarandthedifferenceinmechanicalpropertiesbecomes

lesssignificant

Fig.9 showsaparityplotofFig.9aandthey-axisisrepresented

as‘slurryresistance’.Thevaluesarecalculatedbasedon1/SDRfor

eachresin.Thetrendobservedisapositivepolynomialtrendfor

slurryresistanceagainstcriticalvelocity.Thefirst5resinsshowa

gradualincreaseinmechanicalresistance,however,thedifference

inmechanicalstrengthbecomesmore apparentwhenthemore

robustresins(MabSelectandCaptoQ)areplotted.Thisplot

bet-terdemonstratesthedisparityinmechanicalbehaviourbetween

eachresinanddepictingthedatainthiswaycorrelatespositively

withthepressure-flowdataandismoreeasilycomparable.Both

plotsshowthatDMAcanbeusedasacombinatorytechniqueto

pressure-flowforthecharacterisationofchromatographymedia

DMAhasshownadditionalbenefitsinitsuseforresin

charac-terisation.Itallowsfortheuseofsmallquantitiesofsample(∼1ml)

andrequiresrelativelylittlepreparation.Thelowforceof100mN

appliedina sinusoidalmanneris non-destructivetothemedia

overanextendedperiodoftime(80min).Giventheseadvantages,

itcanpotentiallybeusedtoinvestigatethemechanical

proper-tiesof othermediatypes, e.g.non-agarosebasedresins.It may

alsobeusedinthedevelopmentofnewresinsforrapidtesting

post-emulsification

4 Conclusion and potential applications

Currently,manufacturersusethepressure-flow

characterisa-tiontechniquetodeciphermechanicallimitsofchromatography

mediabypackingcolumnsuptomanylitresinsize.Thisisnot

onlycostly,butitisalsotime-consuminginitspreparation and

itrequiresanumberofbufferstobeused.Wehavedevelopeda

DMAtechniquethatdoesnotrequiretheuseofmultiplebuffers

and uses a much reduced quantityof resin The development

ofthistechniqueconsideredanumberoffactorsalsoassociated

withpressure-flowcharacterisation,includingbeadsize,poresize

and slurry concentration This technique was tested on seven

resinswithvaryingmechanicalpropertiesandcomparedtotheir

pressure-flowcharacteristics.Theresultsshowa strong

correla-tionbetweenbothtechniques.Usingthepressure-flowmethod,the

mostrobustresin(CaptoQ)hadacriticalvelocity3.3timeshigher

thanSepharoseCL-4B,whilsttheDMAtechniqueshowedCaptoQ

tohaveaslurrydeformationrate8.3timeslowerthanSepharose

CL-4B.This couldbe duetoincreasedsensitivity of mechanical

changesasthesamplevolumeusedforDMAismuchsmallerthan

thatofpressure-flow.ThiscorrelationindicatesthatDMAcanbe

usedasacombinatorytechniquefordeterminingmechanical

per-formanceofa givenresin.Althoughadditionaltestscanalways

beperformedtoincrease confidence in itsapplication toother

mediatypes,theresultsfromthisstudyshowdefinitive correla-tionsbetweenthetwo techniquesforagarose-basedresins.The correlationfurthersuggeststhatDMAmaybeappliedtopredict pressure-flowcharacteristics.Thistechniquemayalsobeuseful forrapidtestingofarangeofresinspost-emulsificationandduring thedevelopmentofnewresins.Furthermore,itmayalsobeused

totestresinsexposedtodifferentconditionsinthecolumnaswell

asatdifferentstagesofitslifetimeduringbioprocessing.Itmay alsobeconsideredtoinvestigatetheimpactofexposuretovarying mechanicalstressesduringoperationoflarge-scale chromatogra-phy

Acknowledgments

This work was supported by theEPSRC, Eli Lilly & Co., UCL EastmanDentalInstituteandUCLDepartmentofBiochemical Engi-neering.Thesupportisgratefullyacknowledged

References

[1] N Ioannidis, Manufacturing of Agarose-based Chromatographic Media with Controlled Pore and Particle Size, University of Birmingham, 2009.

[2] G Jagschies, G Sofer, L Hagel, Handbook of Process Chromatography – Development, Manufacturing, Validation and Economics, 2nd ed., Elsevier, 2008.

[3] M Andersson, Characterisation of Chromatography Media Aimed for Purification of Biomolecules, Uppsala Universitet, 2014.

[4] A Jungbauer, Chromatographic media for bioseparation, J Chromatogr A

1065 (1) (2005) 3–12.

[5] A Gerstein, Molecular Biology Problem Solver: A Laboratory Guide, 2nd ed., Wiley, 2004.

[6] Y Mu, A Lyddiatt, A.W Pacek, Manufacture by water/oil emulsification of porous agarose beads: effect of processing conditions on mean particle size, size distribution and mechanical properties, Chem Eng Process Process Intensif 44 (10) (2005) 1157–1166.

[7] Q.-Z Zhou, L.-Y Wang, G.-H Ma, Z.-G Su, Multi-stage premix membrane emulsification for preparation of agarose microbeads with uniform size, J Membr Sci 322 (1) (2008) 98–104.

[8] Q.-Z Zhou, G.-H Ma, Z.-G Su, Effect of membrane parameters on the size and uniformity in preparing agarose beads by premix membrane emulsification, J Membr Sci 326 (2) (2009) 694–700.

[9] Y Yan, Z Zhang, J.R Stokes, Q.Z Zhou, G.H Ma, M.J Adams, Mechanical characterization of agarose micro-particles with a narrow size distribution, Powder Technol 192 (1) (2009) 122–130.

[10] N Ioannidis, J Bowen, A Pacek, Z Zhang, Manufacturing of agarose-based chromatographic adsorbents – effect of ionic strength and cooling conditions

on particle structure and mechanical strength, J Colloid Interface Sci 367 (1) (2012) 153–160.

[11] K Keller, T Friedmann, A Boxman, The bioseparation needs for tomorrow, Trends Biotechnol 19 (11) (2001) 438–441.

[12] A Ellis, J.C Jacquier, Manufacture and characterisation of agarose microparticles, J Food Eng 90 (2) (2009) 141–145.

[13] P Gavara, N Bibi, M Sanchez, M Grasselli, M Fernandez-Lahore, Chromatographic characterization and process performance of column-packed anion exchange fibrous adsorbents for high throughput and high capacity bioseparations, Processes 3 (1) (2015) 204–221.

[14] J.J Stickel, A Fotopoulos, Pressure-flow relationships for packed beds of compressible chromatography media at laboratory and production scale, Biotechnol Progr 17 (4) (2001) 744–751.

[15] R Tran, J.R Joseph, A Sinclair, D Bracewell, Y Zhou, N.J Titchener-Hooker, A framework for the prediction of scale-up when using compressible chromatographic packings, Biotechnol Progr 23 (2) (2007) 413–422.

[16] T Warner, S Nochumson, Rethinking the economics of pensions, BioPharm Int 22 (3) (2003) 15–39.

[17] L Moroni, J.R De Wijn, C.A Van Blitterswijk, 3D fiber-deposited scaffolds for tissue engineering: influence of pores geometry and architecture on dynamic mechanical properties, Biomaterials 27 (7) (2006) 974–985.

[18] T Meyvis, B Stubbe, A comparison between the use of dynamic mechanical analysis and oscillatory shear rheometry for the characterisation of hydrogels, Int J Pharm 244 (2002) 163–168.

[19] M.C Nweke, M Turmaine, R.G McCartney, D.G Bracewell, Drying techniques for the visualisation of agarose-based chromatography media by scanning electron microscopy, Biotechnol J 12 (2016) 1600583.

[20] P DePhillips, A Lenhoff, Pore size distributions of cation-exchange adsorbents determined by inverse size-exclusion chromatography, J Chromatogr A 883 (2000) 39–54.

[21] C Tarmann, A Jungbauer, Adsorption of plasmid DNA on anion exchange chromatography media, J Sep Sci 31 (14) (2008) 2605–2618.

[22] A.M Hardin, C Harinarayan, G Malmquist, A Axen, R van Reis, Ion exchange

chromatography of monoclonal antibodies: effect of resin ligand density on

dymanic binding capacity, J Chromatogr A 20 (1216) (2009) 4366–4371.

[23] Q.-Z Zhou, L.-Y Wang, G.-H Ma, Z.-G Su, Preparation of uniform-sized

agarose beads by microporous membrane emulsification technique, J Colloid

Interface Sci 311 (1) (2007) 118–127.

[25] G.E Healthcare, Packing MabSelect TM and MabSelect SuRe TM Media Using

Verified Methods, GE Healthcare, Uppsala, Sweden, 2015, pp 110.

[26] J Avallin, A Nilsson, M Asplund, N Pettersson, T Searle, C Jägersten,

Columns Upto 1600 Mm in Diameter Packed with Protein A Chromatography

Medium Using Axial Mechanical Compression, GE Healthcare, Uppsala, Sweden, 2016, pp 291.

[27] M Dorn, F Eschbach, D Hekmat, D Weuster-Botz, Influence of different packing methods on the hydrodynamic stability of chromatography columns,

J Chromatogr A (2017).

[28] L Hagel, M Ostberg, T Andersson, Apparent pore size distributions of chromatography media, J Chromatogr A 743 (1) (1996) 33–42.

[29] M Barrande, et al., Characterisation of porous materials for bioseparation, J Chromatogr A 1216 (41) (2009) 6906–6916.