Orthogonal pre-use and post-use efficiency testing for single-use anion exchange chromatography

Three efficiency tests for single-use AEX chromatography devices have been developed and applied to six capsule formats of a new, salt tolerant, single-use AEX product. All the tests have been designed to be performed with simple equipment and common reagents.

Trang 1

Contents lists available at ScienceDirect

journal homepage: www.elsevier.com/locate/chroma

Jonathan F Hestera, ∗, Xinran Lua, Jacob D Calhouna, Rebecca A Hochsteina, Eric J Olsonb

a 3M Separation and Puriﬁcation Sciences Division, 3M Center 236-1C-14, St Paul, MN 55144-10 0 0, United States

b 3M Corporate Research Analytical Laboratory, 3M Center 201-BS-03, St Paul, MN 55144-10 0 0, United States

a r t i c l e i n f o

Article history:

Received 8 April 2021

Revised 12 July 2021

Accepted 22 July 2021

Available online 30 July 2021

Keywords:

Single-use chromatography

Anion exchange (AEX) chromatography

Polishing chromatography

Monoclonal antibodies (MAbs)

Downstream processing

Viral clearance

a b s t r a c t

Threeeﬃciency testsfor single-useAEX chromatographydeviceshavebeen developedand appliedto six capsule formatsof anew, salttolerant,single-use AEX product Allthe testshave been designed

tobeperformedwithsimpleequipmentand commonreagents.Byperformingeachofthethreetests

onundamagedcapsulesandcapsulesintentionallydamagedwithsmalldefects,intandemwithPhi-X174 challengesinahigh-saltbuffer,relationshipsbetweentestresultsandviralclearancehavebeenobtained

A pre-usepressure-basedinstallationveriﬁcation test is simply performedduringequilibration ofthe deviceandeffectiveatidentifyinggrossbypassdefects,forexample,duetointernalsealbreakage.Passing outcomesofapost-useinstallationvalidationbubblepoint testareassociatedwith≥ 5 logreduction value(LRV)ofviralclearance.Anew,non-destructive,pre-useAEXcapacitytestinvolveschallengingthe devicewithchlorideionsandisorthogonaltotheothertwotestsinthatitcandetectchemicaldefects,

as wellas mechanical ones.Passingoutcomes ofthistest correspondto> 2 LRVviral clearanceand providein situassuranceoftheexpectedAEXdynamiccapacitypriortouse.Selectionofapairof pre-useandpost-usetestscanproviderobustriskreductionwithrespecttoviralclearancebysingle-useAEX devicesinbiopharmaceuticalpuriﬁcations

1 Introduction

Flow-through anion exchange (AEX) chromatography is fre-

quently used in biopharmaceutical puriﬁcation processes for re-

duction of net-negatively charged host cell proteins (HCPs) and vi-

ral reduction as part of a validated viral clearance strategy [ 1, 2]

AEX column chromatography is the technology most often used for

electrostatic viral clearance, particularly in commercial scale bio-

pharmaceutical manufacturing, where columns have established a

long history of reliable and well understood performance [3] Still,

validation of HCP and viral clearance by AEX columns in biophar-

maceutical processes involves complexities which contribute sig-

niﬁcantly to operational and regulatory costs Manufacturers must

be concerned with the possibility of micro-channeling in columns

which may result from defects in column packing, a concern rou-

tinely mitigated by in situ measurement of the asymmetry of the

elution peak resulting from the upstream pulse injection of an an-

alyte and quantiﬁcation of the height equivalent to a theoretical

∗ Corresponding author

E-mail address: jfhester@mmm.com (J.F Hester)

plate (HETP) derived from the elution peak retention time and breadth [4–6] Another concern is the potential for loss of viral clearance with resin re-use, which may extend over hundreds of use cycles with intervening cleaning procedures that have the potential to cause resin degradation [6–8] An assessment of the ef- fect of resin re-use on viral clearance is thus generally recommended on a product-by-product basis [ 1, 6, 7]

In recent years, the introduction of single-use AEX technolo- gies has illuminated the potential for reduced regulatory and operational costs associated with flow-through AEX chromatography [ 5, 9–11] Physically resembling and operated like filters, single- use AEX products benefit from improved specific capacity and en- hanced flow rates compared with columns due to the replacement

of diffusive kinetics with convective ﬂow These features have led researchers to note the potential for simpler operation, decreased processing times, and reduced buffer consumption leading to improved economics relative to columns Additionally, their single- use nature obviates validation costs associated with cleaning and performance over repeat use cycles, including viral clearance performance While single-use AEX products were initially used primarily in laboratory and process development activities, recent ad-

https://doi.org/10.1016/j.chroma.2021.462445

( http://creativecommons.org/licenses/by-nc-nd/4.0/ )

Trang 2

vances in ligand and media design have resulted in devices with

high speciﬁc capacity and robust performance across ﬂuid condi-

tions, making them viable alternatives to packed columns in com-

mercial manufacturing [12]

With respect to the deployment of single-use AEX devices in

large-scale manufacturing, investigators have noted the need for

sensitive, in situ test methods capable of detecting device defects

that might result in reduced viral clearance HETP and peak asym-

metry studies used successfully on columns are not sensitive for

single-use devices, given their high ﬂow rates, large mixing vol-

umes, and short bed heights [13] In deﬁning test strategies for

single-use devices, it is useful to consider the viral clearance risks

that might occur Viral clearance loss could occur due to mechan-

ical bypass resulting from media or seal damage which might

originate in manufacturing or during shipping and handling, for

example Alternatively, premature viral breakthrough could occur

in a mechanically integral capsule due to any of a few possible

chemical defects: regions of the media not properly chemically

functionalized during manufacturing or missing AEX media layers,

for example Further, there is a time distribution of risk during

bioprocessing To reduce the risk of processing valuable product-

containing ﬂuid with a damaged capsule, resulting in a process

deviation and the need for costly re-processing, a non-destructive

pre-use eﬃciency test is desired However, there is also a risk that

capsule damage could occur during processing, for example, due to

over-pressurization To mitigate this risk, a post-use eﬃciency test

might be preferred

A number of effort have been made to develop and employ

in situ eﬃciency tests for single-use chromatography devices Dif-

fusion tests are commonly used, wherein the capsule media is

prewet with an aqueous solution, a constant upstream gas pres-

sure is applied, and the gas diffusion rate across the media is

measured [ 14, 15] These are useful for detecting gross mechani-

cal leaks such as seal damage or holes in the media They would

fail to detect chemical defects, however Multiple in situ tests have

been described that can characterize both mechanical and chem-

ical integrity of single-use devices Bind-and-elute type tests in-

volve loading of the AEX device with an excess of a binding an-

alyte, washing the media, and then eluting the analyte and char-

acterizing the resulting elution proﬁle [ 16, 17] Previously described

breakthrough-type tests involve challenging the device with a spe-

ciﬁc analyte while monitoring the concentration of the analyte in

the ﬁltrate Investigators have developed breakthrough-type tests

based on specialized analytes [18]as well as controlled pH changes

in the feed solution [19–23]

In situ eﬃciency tests developed to date generally have one or

more of the following challenges with respect to application at

commercial scale Many of the tests, particularly those of the bind-

and-elute type, rely on subjective comparison of the analyte elu-

tion proﬁle with that of a reference device Quantitative determina-

tion of what deviation from the reference proﬁle is indicative of a

defect expected to result in viral clearance loss, vs. inconsequential

media variability, is diﬃcult, and such subjectivity is not generally

suitable to a Current Good Manufacturing Practices (cGMP) envi-

ronment Some of the tests utilize specialized reagents, not gen-

erally available in biopharmaceutical manufacturing plants, which

would need to be inventoried speciﬁcally for the tests Finally, the

breakthrough-type tests rely on knowledge of the upstream ﬂuid

volume between the ﬂuid injection point and the downstream

breakthrough detector This includes the ﬂuid volume within the

AEX device as well as any upstream and downstream tubing While

such holdup volume estimates are relatively straightforward and

commonly used on lab-scale equipment, they are problematic at

manufacturing scales, where long lengths of large-diameter tubing

may be utilized and where the tubing and capsule headspaces may

contain variable volumes of air bubbles, for example

Herein are described three in situ efficiency tests for single-use AEX devices from which may be selected an orthogonal set of pre- and post-use tests suitable for commercial manufacturing Perfor- mance of the efficiency tests is assessed in the context of a re- cently commercialized, hybrid AEX device that combines a quater- nary ammonium (Q) functional nonwoven AEX media with a novel guanidinium (Gu) functional AEX membrane, achieving robust HCP and virus reduction over wide ranges of fluid pH and ionic strength [12] The test methods have objective and quantitative pass/fail criteria and utilize reagents commonly available in biopharmaceutical manufacturing environments A non-destructive, pre-use pressure-based efficiency test is conveniently performed during equilibration of the device and is effective in identifying gross mechanical defects ( e.g., internal seal damage) that might result

in media bypass A new breakthrough-type, pre-use AEX dynamic capacity test [24] can detect ﬁne mechanical or chemical defects, as well as provide in situ measurement of AEX capacity, without the need to estimate holdup volume of the system Fi- nally, a post-use bubble point test identiﬁes any mechanical defect larger than the largest pore size of the AEX membrane pore size distribution

We present performance data for each of the three eﬃciency tests on undamaged single-use devices of a variety of sizes, as well

as devices with purposefully introduced defects To facilitate quantitative risk assessment, we present data relating eﬃciency test outcomes with measured Phi-X174 viral clearance

2 Experimental

2.1 Materials 2.1.1 Reagents

Sodium acetate (ACS, anhydrous), potassium chloride (ACS), and Tris base (Biotechnology Grade) were purchased from VWR Sodium chloride (U.S.P.) was purchased from J T Baker 1 N hy- drochloric acid solution for buffer pH adjustment was purchased from J T Baker Filling solution for the chloride ion selective electrode (ISE) was purchased from ThermoFisher (Orion ionplus Op- timum Results B, item no 90 0 062) Filling solution for the potassium ISE was purchased from ThermoFisher (Orion ionplus Opti- mum Results E, item no 90 0 065) All reagents were used as received All salt solutions were prepared using deionized water provided by a Millipore Milli-Q water puriﬁcation system which had a resistivity > 18 M •cm

2.1.2 Test capsules

3M TM Polisher ST single-use AEX capsules (3M Company, St Paul, MN) are available in six sizes (denoted BC1, BC4, BC25, BC170, BC340, and BC1020, where the numeral in each designation refers

to the nominal media frontal surface area), as detailed elsewhere [12] The capsules contain a hybrid AEX media bed consisting of 4 layers of Q-functional nonwoven, having a combined bed depth of 0.35 cm, and 3 layers of Gu-functional membrane having a combined bed depth of 0.14 cm Eﬃciency tests were performed on all six capsule sizes

Some pre-use AEX dynamic capacity tests were performed using a 2-piece threaded polycarbonate test housing that could restrain a variable number of 25 mm diameter media discs The housing comprised a top piece with a ﬂuid inlet and a vent for removing air from the housing and a bottom piece with a ﬂuid outlet To simulate the media arrangement of a 3M TM Polisher ST single-use capsule, 3 layers of Gu-functional membrane (FM, 3M Company, St Paul, MN) were placed in the downstream portion

of the housing A plastic annular ring with an inner diameter of

19 mm and an outer diameter of 25 mm was placed on top of

Trang 3

those Next, 4 layers of Q-functional nonwoven (FNW, 3M Com-

pany, St Paul, MN) were placed on top of the ring A second plas-

tic annular ring was placed on top of the 4 layers of FNW Finally,

an o-ring was placed on top of the media stack and the top por-

tion of the housing was screwed down to restrain the media stack

The resulting capsule had an effective frontal media surface area of

2.84 cm 2

2.2 Phi-X174 preparation, ﬁltration challenges, and enumeration

To prepare viral challenge solutions, a concentrated, ﬁlter-

sterilized Phi-X174 virus stock containing 1 × 10 10 plaque-

forming units (PFU)/mL was spiked to a target concentration of

> 1 × 108 PFU/mL into 50 mM Tris-HCl buffer, pH 8.0, adusted

to a conductivity of 20 mS/cm with NaCl After ﬂushing capsules

with 50 L/m 2 of Tris-HCl, pH 8.0, 20 mS/cm buffer at an area-

normalized ﬂow rate of 1 mL/(cm 2-min), the capsules were chal-

lenged with 100 L/m 2of the spiked virus solution at the same ﬂow

rate

The input virus challenge solution and the ﬁltered outputs were

enumerated using a plaque assay Samples were serially diluted to

a concentration at which countable virus plaques (zones of clear-

ing in a bacterial lawn caused by viral lysis) could be visualized

on a 100 mm agar plate (approximately 20–200 plaques/100 μl)

100 μL of the serially diluted virus sample and 50 μl of E coli

13706 (ATCC) overnight host culture were mixed with 2.5 mL of

molten top agar (nutrient broth with 0.5% NaCl and 0.9% agar) and

poured on top of a 100 mm nutrient agar plate Plates were incu-

bated at 37 °C for 3-4 h and plaques were counted The LRV was

calculated by taking the log of the input virus concentration minus

the log of the output concentration Viral clearance results mea-

sured by this assay were expected to be accurate within approxi-

mately ±0.5 LRV [25]

2.3 Pre-use pressure-based installation veriﬁcation test

For BC170, BC340, and BC1020 capsules, a non-destructive, pre-

use, pressure-based installation veriﬁcation test was performed

as speciﬁed by the manufacturer [26] Brieﬂy, the pressure drop

across the capsule was monitored as it was ﬂushed with 20 mM

sodium acetate at an area-normalized ﬂow rate of 600 L/(m 2-h)

A pressure drop greater than or equal to 3 psid was considered

a “pass,” while a pressure drop less than 3 psid was considered

a “fail” possibly indicating a bypass within the capsule To facili-

tate higher-rate testing for BC1, BC4, and BC25 capsules, the pres-

sure drop was monitored as the capsule was ﬂushed with 20 mM

sodium acetate at an area-normalized ﬂow rate of 1,800 L/(m 2-h)

The measured pressure drop was then divided by 3 and the result-

ing value was compared with the ≥ 3 psid test criterion (Pressure

drop was found to be linear with ﬂux within this range of ﬂow

rates.)

2.4 Pre-use AEX dynamic capacity test

2.4.1 Pre-use AEX dynamic capacity tests on capsules of all sizes

using a PendoTECH system

A non-destructive, pre-use eﬃciency test was conducted on

capsules by standardizing the AEX media with bound acetate

counter-ions and then challenging the capsules with dilute potas-

sium chloride [24] While, in the case of an undamaged single-

use capsule, this test fundamentally measures the charge den-

sity of the AEX media within, it is termed a “dynamic capac-

ity” test because it is advantageously performed under ﬂow con-

ditions identical to those used when the capsule performs viral

clearance in use as recommended by the manufacturer Fig.1 is

a schematic of a tabletop setup that was used to perform this

pre-use AEX dynamic capacity test on all 3M TM Polisher ST capsule formats A peristaltic pump (Cole-Parmer item no ZM-07522-

20 with Masterﬂex L/S Easy-Load II Pump Head, item no 77200- 60) supplied any of 1 M sodium acetate, 20 mM sodium acetate,

or 20 mM KCl solution to the capsule inlet, with valves posi- tioned upstream of the pump as shown to select the challenge solution For BC1, BC4, and BC25 capsules, potassium and chloride ISE’s were mounted in a dual-probe plexiglass ﬂow cell (Item No

791101, FIAlab Instruments, Seattle, WA) with the potassium ISE upstream of the chloride ISE For larger BC170, BC340, and BC1020 capsules, the ISE’s were mounted in a purpose-built polycarbonate ﬂow cell with top and bottom pieces that were joined together and sealed with an o-ring, deﬁning a small internal volume in contact with the ISE tips Upstream and downstream pressure transducers were monitored using a pressure monitor/transmitter (PMAT2A, PendoTECH, Princeton, NJ) The ISE outputs were monitored using

a pH/conductivity monitor/transmitter (PDKT-ACCESS-PHCN, Pen- doTECH) Filtrate mass was monitored using a top-loading scale Mass, ISE output, and pressure data were automatically logged using PendoTECH PressureMAT data logging software (PMATP-GUI, PendoTECH)

The capsule was initially ﬁlled with 1 M sodium acetate solution, with the vent open, until all air was removed from the capsule, after which the vent was closed With reference to Fig.1, in

a first step, bypass valve 6 was directed to bypass and the capsule was flushed with 1 M sodium acetate to standardize the media by replacing all bound counter-ions with acetate In a second step, the capsule was washed with 20 mM sodium acetate to remove from the capsule housing the high acetate concentration from the pre- vious step In a third step, bypass valve 6 was set to direct the filtrate flow through the ISE flow cell and the flow rate of the 20 mM sodium acetate feed solution was reduced to equilibrate the system

at the ﬂow conditions of the subsequent challenge step In a fourth step, the capsule was challenged with a 20 mM potassium chloride solution while the downstream ISE responses were monitored Test conditions for each capsule format are detailed in Table1 Potassium ions are unbound by the AEX media and pass through the capsule with the challenge ﬂuid front Chloride ions are bound by the media, displacing acetate ions Fig.2is a plot of exemplary resulting sets of potassium and chloride ISE responses for three different 3M TM Polisher ST BC340 capsules The oppos- ing directionality of the K + and Cl −breakthrough responses is ex- plained by the Nernst equation, wherein the response of an elec- trochemical cell to an ionic analyte is proportional to the logarithm

of the analyte concentration multiplied by a term including the charge of the analyte Assuming collection of the ﬁltrate during the challenge step begins at the moment the feed solution is switched from 20 mM sodium acetate solution to the KCl challenge solution, the potassium ion breakthrough volume is the system holdup volume between the feed solution injection point and the ISE’s The chloride ion breakthrough volume is the volumetric throughput at which the AEX capacity of the device is exhausted The AEX capacity of the device is characterized by the net breakthrough volume ( Vnet), the difference between the chloride and potassium breakthrough volumes This value was normalized by the capsule surface area and expressed in terms of volume of ﬁltrate per unit area (1909 mL / 340 cm 2 = 5.6 mL/cm 2 for capsule 3 in Fig 2)

or microequivalents of chloride per unit area based on the 20 mM challenge solution concentration (112 μequiv/cm 2 for capsule 3 in Fig.2) A “pass” criterion of ≥ 4 mL/cm 2 ( ≥ 80 mequiv/cm 2) was selected on the basis of numerous trials conducted on capsules containing Q-functional nonwoven and Gu-functional membrane media spanning the media speciﬁcation ranges for AEX capacity Thus, undamaged capsules containing media of both types at the lower speciﬁcation limit for AEX capacity are expected to pass the above criteria

Trang 4

Fig 1 Schematic diagram of setup used to perform pre-use AEX capacity tests on all capsule formats Solid lines denote ﬂuid (tubing) connections, while dashed lines

denote electrical connections

Fig 2 Exemplary pre-use AEX capacity test data for three 3M TM Polisher ST BC340 (340 cm 2 ) capsules [24] Open symbols are potassium ISE responses (left axis) while closed symbols are chloride ISE responses (right axis)

Trang 5

Table 1

Steps and settings for performing the pre-use integrity test on any capsule size using a PendoTECH control and data acquisition system Numerals in black circles reference the corresponding labels in Fig 1

End criteria End when signiﬁcant Cl − ISE breakthrough response is observed

2.4.2 Pre-use AEX capacity tests on small capsules utilizing a Cytiva

ÄKTA chromatography system

Pre-use AEX dynamic capacity tests on media conﬁgurations

in the polycarbonate test housing were performed using a Cytiva

ÄKTA avant 25 chromatography system (Item No 28930842, Cy-

tiva) Potassium and chloride ISE’s were mounted in a dual-probe

plexiglass ﬂow cell (Item No 791101, FIAlab Intsruments, Seattle,

WA) with the potassium ISE upstream of the chloride ISE The re-

sulting ISE probe assembly was mounted on the ÄKTA chromatog-

raphy system with HPLC tubing from the “Out1” position of the

ÄKTA outlet valve directed to the inlet of the side of the ﬂow cell

containing the potassium ISE and tubing from the opposite side

of the ﬂow cell directed to a waste container The ÄKTA system

was outﬁtted with a Cytiva I/O-box E9 (Item No 29011361, Cytiva)

to enable automatic data logging of the ISE outputs Input to the

I/O-box E9 was provided by two analog cables with female jacks

(Item No 290-1009-ND, Digi-Key) The ISE’s were connected to the

input cables using two BNC female interconnect jacks (Item No

ARF1069-ND, Digi-Key) Automated data collection from the ÄKTA

system was provided by Cytiva UNICORN software, which was con-

ﬁgured to log I/O-box E9 inputs according to the manufacturer’s in-

structions Deionized water was setup on system pump “A” at posi-

tion “A1.” A 0.5 M sodium acetate equilibration solution was setup

on system pump “B” at position “B1.” The 20 mM KCl challenge so-

lution was setup on the sample pump at position “S1.” Generally,

Cytiva ÄKTA avant and ÄKTA pure chromatography system models

are suitable for performing the pre-use AEX dynamic capacity test

on 3M TM Polisher ST BC1, BC4, and BC25 capsules

The polycarbonate housing was assembled as described in

Section 2.1.2 and the resulting ﬁlter assembly was mounted on

a column position of the ÄKTA system With the vent open, 0.5

M sodium acetate solution from inlet position “B1” was pumped

slowly into the housing until all the air was removed and then

the vent was closed The media was then equilibrated by pumping

24 mL of 0.5 M sodium acetate through the housing at a ﬂow rate

of 12 mL/min with the ﬁltrate directed to outlet valve position “W” (waste) The ÄKTA system was then conﬁgured to deliver a 20 mM sodium acetate wash solution as a gradient comprising 4% of 0.5

M sodium acetate solution from inlet position “B1” and 96% deionized water from inlet position “A1.” 12 mL of the wash solution was pumped through the housing at a flow rate of 6 mL/min with the filtrate directed to outlet valve position “W.” Then, the outlet valve position was switched to “Out1,” to direct the filtrate to the ISE assembly, and 12 mL of the 20 mM sodium acetate wash solution was pumped through the housing and ISE assembly at a flow rate of 3 mL/min Finally, the injection valve was moved to change the inlet flow source to the 20 mM KCl challenge solution at sample pump position “S1,” and the challenge solution was pumped through the capsule at 3 mL/min The potassium and chloride ISE responses were monitored, and flow was continued until complete breakthrough responses had been detected by both probes For each run, the filtrate volume at the initial inflection point of each of the potassium and chloride breakthrough responses was recorded as the corresponding breakthrough volume, and the AEX capacity was calculated as described in Section2.4.1

2.5 Post-use installation validation bubble point test

A post-use installation validation bubble point test was performed on capsules as speciﬁed by the manufacturer [26] Brieﬂy, the capsule headspace was gravity drained and then the inlet was connected to a Sartocheck® 4 Plus Filter Tester (Sartorius) while

a length of tubing connected to the capsule outlet was immersed

in water An upstream air pressure of 50 mbar was applied, and the air pressure was ramped up in 50 mbar increments until the bubble point pressure of the capsule was detected by visual obser-

Trang 6

vation of a stream of bubbles emerging from the outlet tubing A

bubble point pressure greater than or equal to 900 mbar was con-

sidered a “pass,” while a bubble point pressure less than 900 mbar

was considered a “fail” possibly indicating a mechanically defec-

tive capsule Note that, while some single-use AEX device man-

ufacturers recommend similar gas diffusion type tests as pre-use

tests [ 14,15], the bubble point test described above must be con-

sidered destructive for 3M TM Polisher ST and thus cannot be per-

formed as a pre-use test, as it can introduce air between the FNW

layers that may not be reliably removed

2.6 Controlled damage of AEX capsules

To create various controlled levels of media damage, some

3M TM Polisher ST BC1, BC4, BC25, BC170, BC340, and BC1020 cap-

sules were prepared with small, controlled defects by introduc-

ing holes through the full media stack using blunt-tipped nee-

dles (Small Hub RN Needle, Point Style 3, Hamilton, Reno, NV) of

various diameter ranging from 32 gauge (0.11 mm) to 18 gauge

(0.84 mm) In the case of BC1, BC4, and BC25 laboratory capsules,

the holes were introduced in the stacked media prior to capsule

assembly In the case of BC170 capsules, holes were introduced in

the media in the internal lenticular ﬁlter element prior to assem-

bling the capsule In the case of BC340 and BC1020 capsules, holes

were introduced through one entire lenticular element, including

media stacks on both sides of the element, prior to assembling the

capsule One BC170 capsule was assembled using an undamaged

lenticular ﬁlter element but with a purposeful misalignment of an

o-ring that forms a seal between the ﬁlter element and the hous-

ing; this was expected to create a gross mechanical bypass within

the capsule

3 Results and discussion

3.1 Performance and repeatability of the pre-use AEX dynamic

capacity test on integral capsules

Five sequential AEX dynamic capacity tests were performed on

each of three 3M TM Polisher ST BC340 capsules, produced using

the same AEX media lots, as described in Section 2.4.1 One rep-

resentative set of potassium and chloride responses for each cap-

sule appears in Fig 1 Potassium ISE responses featured a grad-

ual, monotonically diminishing value prior to arrival of potassium

ions at the ISE ﬂow cell, followed by the rapid onset of an increas-

ing ISE response which was recorded as the K +breakthrough vol-

ume as illustrated in Fig 1 The internal BC340 capsule volume

of 730 mL [ 27] deﬁnes an expected lower limit of the K + break-

through volume In practice, measured potassium breakthrough

volumes were somewhat larger and variable due to the volume

of tubing between the challenge solution injection point and the

ISE ﬂow cell and also due to the fact that the test operator col-

lected variable volumes of ﬁltrate on the mass balance during the

20 mM sodium acetate equilibration step before switching the feed

to the KCl challenge solution The chloride ISE responses featured

a gradual, monotonically increasing value prior to chloride break-

through, followed by the onset of a monotonically decreasing re-

sponse which was recorded as the Cl − breakthrough volume as il-

lustrated in Fig 1 K + and Cl − breakthrough volumes and calcu-

lated dynamic capacities for all 15 trials are detailed in Table 2

The test was reproducible with a coeﬃcient of variation well be-

low 10 percent

3M TM Polisher ST capsules are designed for single-use AEX

chromatography in a ﬂow-through mode They feature shallow AEX

bed depths relative to columns, and engineering trade-offs have

been made to facilitate additional performance features, such as

relatively large capsule headspaces to enable rapid flow and handling of turbid fluids [12] The large resulting mixing volume upstream of the AEX media is responsible for the broad breakthrough curves relative to columns Selection of acetate as the standardizing anion is important; KCl challenges conducted on capsules stan- dardized with sodium bicarbonate or sodium sulfate featured earlier and broader chloride breakthrough curves and reduced dis- tinction between integral and damaged capsules (data not shown) This might be expected, since acetate has a lower selectivity coeffi- cient with respect to the Q-functional media than chloride, resulting in a self-sharpening ion exchange boundary within the media, whereas more tightly-binding standardizing anions than chloride would be expected to result in an ion exchange boundary spread- ing with progression through the media [28]

Attempts were made to calibrate the ISE’s such that the breakthrough curves could be plotted in units of ion concentration rather than as raw ISE electrical responses as in Fig 1 These attempts were frustrated by signiﬁcant ISE response drift over time originating from a number of factors Among these, both ISE’s are essentially measuring “zero” concentration prior to breakthrough

of their respective analyte ions, the resulting cell potential of which is not well defined This accounts for the variable vertical positions of the breakthrough curves in Fig 1 and other fig- ures herein A second factor is relatively rapid drifts in the quantitative ISE responses over subsequent experiments, for example, due to plasticization of the polymeric membrane of the potassium ISE by acetate [29] Due to these challenges, ISE responses herein are plotted as raw electrical responses, since the primary response attributes of value are the filtrate volumes at which K + and Cl − breakthrough occur

3.2 Detection of missing media layers using pre-use AEX dynamic capacity test

Fig.3is a plot of potassium and chloride ISE responses during the AEX dynamic capacity test for a complete media stack of 4 layers of FNW and 3 layers of FM in a 25 mm test housing (3 replicates), followed by ISE responses for sequential trials in which one layer of media was removed from the housing in each trial The ISE response inﬂection volumes were used to compute volumetric capacities as detailed in Table3

The full media stack exhibits a reproducible and passing ( ≥4 mL/cm 2) volumetric chloride challenge capacity Media stacks with “missing” layers exhibit capacities that fail the test criterion and steadily decrease as media layers are removed While the chloride breakthrough volume shifts signiﬁcantly with the removal of each media layer, the potassium breakthrough volume shifts only slightly, reﬂecting the smaller hold-up volume as the housing volume decreases with layer removal An empty housing exhibits nearly coincident potassium and chloride breakthrough events These results illustrate the capability of the non-destructive AEX dynamic capacity test to detect missing functional media layers and characterize the AEX capacity of single-use capsules prior

to use, and also the hold-up volume independence of the method The potassium ISE response curves exhibit a second, minor in- ﬂection coincident with chloride breakthrough This is a result of the change in chloride concentration at the reference electrode liq- uid junction of the potassium ISE, the internal ﬁlling solution of which is aqueous sodium chloride [30]

3.3 Detection of damaged media using pre-use AEX dynamic capacity test

Fig 4 is a plot of potassium and chloride ISE responses during the AEX dynamic capacity test for an undamaged BC1 capsule and for BC1 capsules containing damaged media stacks, each

Trang 7

Table 2

ISE breakthrough volumes and calculated AEX dynamic binding capacities for 15 trials conducted on three 3M TM Polisher ST BC340 capsules produced

using the same AEX media lots

Capsule Trial K + Breakthrough Volume, mL Cl − Breakthrough Volume, mL Volumetric capacity, mL/cm 2 Cl − capacity, μequiv/cm 2

Fig 3 Plot of chloride ISE response curves (top set of curves) and potassium ISE response curves (bottom set of curves) for 25 mm test housing containing various sets of

AEX functional media layers Data collected using an ÄKTA avant 25 chromatography system

Trang 8

Table 3

ISE breakthrough volumes and calculated AEX dynamic binding capacities for 25 mm test housing containing various sets of AEX functional media layers

Media Construction K + Breakthrough Volume, mL Cl - Breakthrough Volume, mL Volumetric capacity, mL/cm 2 Cl - capacity, μequiv/cm 2

Fig 4 Plot of chloride ISE response curves (top set of curves) and potassium ISE response curves (bottom set of curves) for BC1 (1 cm 2 ) capsules containing media stacks pierced with blunt-tipped needles of various diameter All curves have been shifted on the x -axis such that the potassium breakthrough inﬂection point was set at zero volume Response curves have been intentionally vertically offset for easy comparison Y -axis tick marks are at intervals of 0.1 mA Hole sizes and net breakthrough volumes for each curve are indicated Data collected using the PendoTECH setup

of which was pierced once with a blunt-tipped needle varying in

diameter from 32 gauge (0.11 mm) to 22 gauge (0.41 mm) The

media stack comprises two AEX media with different characteris-

tics with respect to these piercing defects Whereas piercing of the

FM creates a deﬁned, roughly circular hole, the FNW swells upon

wetting in buffer and appears capable of “healing” small puncture

defects upon swelling At the largest hole size, it appears that a

relatively well-deﬁned low-pressure path is created in the media

stack, and chloride breakthrough is observed to be nearly coin-

cident with potassium breakthrough At smaller hole sizes, chlo-

ride breakthrough occurs earlier than for an undamaged capsule

and the breakthrough curves broaden This is consistent with the

superposition of fast chloride breakthrough within a low-pressure path comprising partially “healed” FNW layers and typical chloride breakthrough within the remainder of the media The pre-use AEX dynamic capacity test is quite sensitive to small defects, with a defect as small as 0.003% of the frontal media area having a volumetric capacity <4 mL/cm 2and failing the test criterion

3.4 Relation of pre-use pressure-based test, pre-use AEX dynamic capacity test, and post-use bubble point test results with viral clearance

A total of 111 capsules of various format were made, one of which contained a purposefully misaligned internal seal and many

Trang 9

Fig. 5 Summary plot of observed pre-use pressure-based installation veriﬁcation test values vs measured Phi-X174 viral clearance Upward arrows denote viral clearance

experiments in which no virus plaques were observed after plating of the filtrate, and the minimal viral clearance was thus defined by the measured concentration of the challenge Dashed line highlights the pre-use installation verification test criterion

Table 4

Undamaged and damaged capsules tested in experiments relating eﬃciency test outcomes to viral clearance

Needle Size No of Media Lot Combos / No of Replicates per Lot Combo Gauge Diameter, mm BC1 BC4 BC25 BC170 BC340 BC1020

of which were intentionally damaged with blunt-tipped needles

as detailed in Table 4, to examine the relationship between each

of the three eﬃciency tests, performed on integral and defective

capsules, with Phi-X174 viral clearance in a high conductivity chal-

lenge An effort was made to ensure the capsule population in this

study was representative of the “real world” variability expected

to exist among manufactured capsules For example, the FM me-

dia primarily responsible for AEX viral clearance in these devices,

particularly at high conductivity [12], is released in manufacturing

on the basis of its measured bovine serum albumin (BSA) dynamic

binding capacity (DBC), which has minimum and maximum release

values of 7.9 and 12.0 mg/cm 2, respectively [27] Each of the 111

capsules in this study was built using one of 11 different lots of

FM with BSA DBC spanning nearly the entire allowable range of BSA DBC, from 8.2 to 11.9 mg/cm 2

Each capsule was tested according to the following work- flow First, a pre-use pressure-based installation verification test was conducted using 20 mM sodium acetate during the pre- conditioning flush required by the manufacturer to remove glycerin used to stabilize the functional media Second, a pre-use AEX dynamic capacity test was performed using the PendoTECH setup Third, a Phi-X174 challenge was conducted under high-salt buffer conditions (50 mM Tris-HCl, pH 8.0, adjusted to 20 mS/cm with NaCl) Finally, a post-use installation validation bubble point test was conducted

Trang 10

Fig. 6 Summary plot of observed pre-use AEX dynamic capacity test values vs measured Phi-X174 viral clearance Upward arrows denote viral clearance experiments in

which no virus plaques were observed after plating of the ﬁltrate, and the minimal viral clearance was thus deﬁned by the measured concentration of the challenge Vertical dashed line highlights the pre-use AEX dynamic capacity test criterion Horizontal dashed line highlights a 2-LRV viral clearance level

Figs.5– are summary plots of pre-use pressure-based installa-

tion veriﬁcation test, pre-use AEX dynamic capacity test, and post-

use installation validation bubble point test values, respectively,

versus Phi-X174 LRV for undamaged and needle-damaged capsules

The pre-use pressure-based installation validation test successfully

detected the gross bypass resulting from the misaligned seal in the

BC170 capsule damaged in that way, providing a failing pressure

drop of 0.6 psid as compared to a passing pressure drop of 12.9

psid for an undamaged and properly assembled BC170 capsule

However, the data in Fig.5indicate this simple pressure-based test

is not capable of detecting small defects (like ﬁne needle piercings)

that can result in substantial loss of viral clearance

As shown in Fig 6, the pre-use AEX dynamic capacity test is

more effective at detecting small defects, providing a positive cor-

relation between measured AEX dynamic capacity and viral clear-

ance The correlation is not excellent (linear R 2 = 0.67), which

might be expected based on the manufacturing variability of the

AEX media as noted above, the maximum measurable viral clear-

ance of just greater than 7 LRV deﬁned by the challenge con-

centration, and the expected viral clearance plaque assay accuracy

of about ±0.5 LRV Still, based on the testing of 110 capsules of

all six formats selected to be representative of approximately the

full range of manufacturing variability, passage of the test crite-

rion was associated with greater than 2 LRV viral clearance in a

high-salt buffer Additionally, this test is capable of detecting not

only mechanical defects, but also chemical ones, and provides a

non-destructive pre-use measurement of the AEX capacity of the

single-use capsule

As shown in Fig.7, the post-use installation validation bubble

point test is the most effective test for characterizing small me-

chanical defects In one case, a BC170 capsule with a 28-gauge hole had a post-use bubble point pressure greater than the test criterion and had a measured viral clearance of > 4.84 LRV No virus plaques were observed in the plated viral challenge ﬁltrate, however, and the value of 4.84 LRV was thus a lower limit of the viral clearance because the concentration of the spike in this particu- lar challenge was insuﬃcient to measure > 5 LRV In all other of the 110 capsules tested, passage of the post-use bubble point test criterion was associated with ≥ 5 LRV Phi-X174 clearance

4 Conclusion

Three efficiency tests have been developed and applied to characterize mechanical and chemical defects in six capsule formats of 3M TM Polisher ST, a single-use AEX chromatography product Two non-destructive, pre-use tests can be applied to reduce the risk of processing product-containing fluid with a damaged or defective capsule A pre-use pressure-based installation verification test is conveniently applied, with minimal equipment (only a pump and

a pressure gauge or transducer), during the pre-conditioning ﬂush procedure required to remove glycerin stabilizer from the capsule media and equilibrate the capsule for use This simple test is capable of detecting gross mechanical defects, such as broken internal seals, that might result from shipping damage, for example

A new pre-use AEX dynamic capacity test is somewhat more complicated to execute, though it uses common reagents This test is more effective at reducing the risk of processing product- containing ﬂuid using a capsule containing even very small defects, with passing test results associated with > 2 LRV viral clearance

in a high-salt buffer challenge In addition, the AEX dynamic ca-

Tiêu đề	Orthogonal Pre-use and Post-use Efficiency Testing for Single-use Anion Exchange Chromatography
Tác giả	Jonathan F. Hester, Xinran Lu, Jacob D. Calhoun, Rebecca A. Hochstein, Eric J. Olson
Trường học	3M Company
Chuyên ngành	Biopharmaceutical Purification
Thể loại	Research Article
Năm xuất bản	2021
Thành phố	St. Paul

Định dạng
Số trang	12
Dung lượng	2,08 MB