Application of cation exchange chromatography in bind and elute and flowthrough mode for the purification of enteroviruses

Members of the enterovirus genus are promising oncolytic agents. Their morphogenesis involves the generation of both genome-packed infectious capsids and empty capsids. The latter are typically considered as an impurity in need of removal from the final product.

Journal of Chromatography A 1676 (2022) 463259 Contents lists available at ScienceDirect Journal of Chromatography A journal homepage: www.elsevier.com/locate/chroma Application of cation exchange chromatography in bind and elute and flowthrough mode for the purification of enteroviruses Spyridon Konstantinidis a,∗, Murphy R Poplyk a, Andrew R Swartz a, Richard R Rustandi b, Rachel Thompson b, Sheng-Ching Wang a a b Vaccine Process Research and Development, Merck & Co., Inc., Rahway, NJ, USA Analytical Research and Development, Merck & Co., Inc., Rahway, NJ, USA a r t i c l e i n f o Article history: Received April 2022 Revised 15 June 2022 Accepted 16 June 2022 Available online 17 June 2022 Keywords: Enterovirus Empty capsids Cation exchange chromatography High throughput Oncolytic virus a b s t r a c t Members of the enterovirus genus are promising oncolytic agents Their morphogenesis involves the generation of both genome-packed infectious capsids and empty capsids The latter are typically considered as an impurity in need of removal from the final product The separation of empty and full capsids can take place with centrifugation methods, which are of low throughput and poorly scalable, or scalable chromatographic processes, which typically require peak cutting and a significant trade-off between purity and yield Here we demonstrate the application of packed bed cation exchange (CEX) column chromatography for the separation of empty capsids from infectious virions for a prototype strain of Coxsackievirus A21 This separation was developed using high throughput chromatography techniques and scaled up as a bind and elute polishing step The separation was robust over a wide range of operating conditions and returned highly resolved empty and full capsids The CEX step could be operated in bind and elute or flowthrough mode with similar selectivity and returned yields greater than 70% for full mature virus particles Similar performance was also achieved using a selection of other bead based CEX chromatography media, demonstrating general applicability of this type of chromatography for Coxsackievirus A21 purification These results highlight the wide applicability and excellent performance of CEX chromatography for the purification of enteroviruses, such as Coxsackievirus A21 © 2022 Merck Sharp & Dohme Corp., a subsidiary Merck & Co., Inc., Kenilworth, NJ, USA Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) Introduction Enteroviruses belong to the picornavirus family and are nonenveloped viruses with a diameter of 27–30 nm They have a positive sense single stranded RNA genome of ∼7.4 kb long encoding structural and non-structural proteins necessary for their replication [1,2] Such viruses are attractive oncolytic agents employed in cancer treatment [2–4] Recently, a prototype strain of Coxsack- Abbreviations: Abs., Absorbance; AC, Affinity capture; BSA, Bovine serum albumin; Csalt , Salt concentration in gradient; Csalt,o , Starting salt concentration in gradient; CCCH, Clarified cell culture harvest; CEX, Cation exchange; CV, Column volume; CVelution , Column volume number in the elution phase of a column; CVA, Coxsackievirus; E1 – E3, Elution pools – 3; ED, Effective dose; FT1 – FT5, Flowthrough pools – 5; GSH, Glutathione; HCP, Host cell protein; HT, High throughput; IEX, Ion exchange; PAGE, Polyacrylamide gel electrophoresis; SDS, Sodium dodecyl sulfate; S, Strip pool; sd, Standard deviation; VP0 – 4, Viral polypeptide - 4; W, Wash pool ∗ Corresponding author E-mail address: spyridon.konstantinidis@merck.com (S Konstantinidis) ievirus A21 (CVA21) was also demonstrated as a potentially novel therapeutic for bladder cancer [5] in addition to other cancers involving tumors overexpressing the cell surface receptor intercellular adhesion molecule [6] The capsids of full mature enterovirus virions are composed of 60 copies of four viral proteins (VP) VP1– VP4 arranged in a shell that packages the RNA genome The generation of such full mature virus virions is the result of a complex morphogenesis comprised of multiple steps [7,8] Upon receptor binding and delivery of the RNA genome, the nascently expressed P1 polyprotein is cleaved by virally encoded proteases into VP0, VP1 and VP3 These associate to form protomers ([(VP0, VP1, VP3)1 ]) which are then assembled into pentamers ([(VP0, VP1, VP3)5 ]) Pentamers can assemble into either empty procapsids ([(VP0, VP1, VP3)5 ]12 ) or encapsidate the replicated genome to form provirions ([(VP0, VP1, VP3)5 ]12 RNA)) Provirions undergo an autocatalytic cleavage of VP0 into VP2 and VP4, and the accrued re-arrangement of capsid proteins results into stable icosahedral full mature virions ([(VP1, VP2, VP3, VP4)5 ]12 RNA) https://doi.org/10.1016/j.chroma.2022.463259 0021-9673/© 2022 Merck Sharp & Dohme Corp., a subsidiary Merck & Co., Inc., Kenilworth, NJ, USA Published by Elsevier B.V This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/) S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Salisbury, UK) were cultured on g L−1 Cytodex-1 microcarriers (Cytiva) using GIBCOTM William’s E Medium (Thermo Fisher Scientific Inc., MA, USA), supplemented with 10% (v/v) HyCloneTM Bovine Calf Serum (Cytiva), 0.1% P188, mM L-glutamine, and 20 mM glucose Cell culture took place at controlled conditions of 37 °C and pH 7.2 At days post bioreactor batching, and once the cells had reached >90% confluency, the bioreactor underwent a 80% media exchange into serum-free cell culture media The temperature of the reactor was then controlled to either 37 °C (process A) or 34 °C (process B) Upon media exchange, the cells were infected with a multiplicity of infection of 0.05 using a CVA21 virus stock (cat VR-860) At days post infection, and at an observed cytopathic effect of >90%, the reactor was harvested The collected viral fluid was filtered with a Sartopure® GF+ depth filter (Sartorius, Gưttingen, Germany) and clarified with a Sartoclean® CA μm|0.8 μm filter (Sartorius) before it was processed further Alternatively, the clarified harvest was stored at °C or -70 °C for short and long term storage, respectively, until further testing Results presented hereafter employ CVA21 material generated from process B, unless stated otherwise Virus maturation can be affected by a plethora of factors [8] and in vitro cell culture production of enteroviruses, such as CVA21, may not lead exclusively to the assembly of full mature virions, which are the infectious particles displaying the desired oncolytic activity; instead non-infectious particles may be assembled, such as empty procapsids The latter can potentially elicit undesired immune responses and are often the subject of scrutiny from regulatory authorities [9] The separation of full mature virions from empty procapsids, or full particles/capsids from empty particles/capsids, is a challenging task typically achieved by differential centrifugation which exploits buoyancy density differences between the particles [10,11] Such separations are, however, not desirable for large scale bioprocessing [12,13] Hence, alternative routes for purifying full mature virions from empty procapsids, in addition to process (e.g., host cell proteins) and product related impurities, are sought Empty procapsids are routinely encountered during the processing of adeno-associated virus vectors and here their separation from full particles has been achieved by employing ion exchange chromatography (e.g., [14–16]) However, this separation method results in closely-eluting full and empty particle peaks and requires peak cutting which is challenging to implement at manufacturing scale and can result in yield losses in favor of purity [17] Similarly, chromatography-based purifications of enteroviruses return low product yields and focus predominantly on the reduction of process related impurities [18,19] Recently, a novel glutathione affinity chromatography (GSH AC) capture step has been shown to purify CVA21 from clarified cell culture harvests [20] However, upstream process conditions at infection can challenge this step due to an undesired co-elution of both CVA21 full mature virions and empty procapsids Here, we report the deployment of cation exchange (CEX) chromatography as a polishing step for the purification of CVA21 The virus is produced in adherent cell culture and purified in a three-column process, which employs the GSH AC step, an intermediate ion exchange (IEX) chromatography step, and the CEX polishing step High throughput chromatography techniques were employed to develop the CEX-based polishing step and to generate information regarding its wide applicability It is demonstrated that CEX chromatography can be deployed robustly in either bind and elute or flowthrough mode, returning in both cases mature virions in high yields while eliminating empty procapsids from the resultant product pool When deployed in bind and elute mode, the polishing step eluted full mature virions in a concentrated form and the separation displayed baseline resolution from empty procapsids The scalability of the CEX polishing step was also demonstrated and it was furthermore shown that efficient and effective purification of the CVA21 full particles from empty procapsids could be obtained using a diverse selection of cation exchange resins These results serve to demonstrate the value of cation exchange chromatography in the purification of enteroviruses, such as CVA21 2.2 Large scale GSH affinity column chromatography GSH affinity chromatography (GSH AC) was performed as described in [20] using a GSH Sepharose® FF column Briefly, the column was typically loaded with 200 column volumes (CVs) of clarified cell culture harvest (CCCH) and eluted in CVs with a 15 mM Tris, pH 8.0, 100 mM NaCl, mM Dithiothreitol, mM GSH buffer At large scale, the collected elution pool (i.e., GSH AC product) was purified further with the application of an intermediate IEX step, and a bind and elute CEX step Conversely, at high throughput (HT) scale, the GSH AC product was employed directly for the development of the CEX step Purified intermediates were stored at °C or –70 °C for short and long term storage, respectively All buffers contained 0.005% polysorbate-80 (PS80) The GSH AC capture of CVA21 produced from upstream process B, led to a co-elution of full and empty CVA21 particles Application of this step to CVA21 material generated from upstream process A led to reduced empty procapsids in the elution product pool 2.3 High throughput chromatography 2.3.1 RoboColumn chromatography Miniature column HT chromatography experiments employed Opus® RoboColumns® (Repligen, MA, USA) on a Tecan EVO® 150 robotic station (base unit), which was equipped with an 8–channel liquid handling arm and an eccentric robot manipulator arm and operated by EVOware® v2.8 The station was fitted with short stainless-steel tips and integrated with Te-ChromTM , Te-ShuttleTM , and Infinite® M10 0pro reader devices The described configuration allowed for up to eight chromatographic separations to be executed in parallel in a process described in [21] Here, all buffers and solutions contained 0.005% PS80, and all chromatography phases were run at a residence time of A total of 16 RoboColumn-based separations were performed using GSH AC product as feed (Table 1) Separations #1–14 and #16 employed 200 μL RoboColumns and they aimed to evaluate the separation of full mature virions from empty procapsids on a selection of ion exchange resins Separation #15 employed 600 μL RoboColumns and sought to evaluate the separation of full mature virions from process related impurities Furthermore, separations #1–6 and #10–16 were run in bind and elute mode, whereas separations #7–9 were run in flowthrough mode For all separations, fractions were collected every 200 μL in fullarea UV transparent 96 well microplates (Corning Life Sciences) and read on the M10 0pro reader at 260 nm, 900 nm and 990 Materials and methods In this work, unless specified otherwise, chemicals and buffers were from Sigma-Aldrich® (MO, USA), chromatography resins, columns, stations, and 96 well PreDictorTM plates were from Cytiva (Uppsala, Sweden), robotic stations and pertinent components were from Tecan Group Ltd (Männedorf, Switzerland), and virus stocks were acquired from ATCC (VA, USA) 2.1 Material generation 2.1.1 Generation of CVA21 enterovirus CVA21 was produced using L BioFlo® vessels operated through a BioFlo 320 bioprocess control station (Eppendorf, NY, USA) Human lung fibrobast MRC-5 cells (cat 05072101; ECACC, S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Table Details of high throughput RoboColumn chromatography separations carried out to screen the polishing purification of Coxsackievirus A21 via bind and elute and flowthrough mode cation exchange chromatography and anion exchange chromatography as a function of resin, mobile phase conditions, phase duration, in column volumes (CVs), and gradient slope In all separations the pH value was kept constant across all phases other than the stripping of the columns [NaCl] depicts the NaCl concentration in the equilibration (Equil.), load, and wash buffer in each separation This concentration was also the starting concentration in the elution gradient where applicable Separations #1–6, #10–14, and #16, employed 200 μL RoboColumns and clarified cell culture harvest (CCCH) from upstream process B Separation #15 employed 600 μL RoboColumns and CCCH from upstream process A In all separations, the collected fractions had a nominal volume of 200 μL Separation #1 #2 #3 #4 #5 #6 #7 #8 #9 #10 #11 #12 #13 #14 #15 #16 Resin Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Poros 50 HS Capto S ImpAct Capto SP ImpRes Capto S Nuvia HR-S Nuvia S Poros 50 HS Nuvia HP-Q pH 3.8 4.0 4.2 4.5 5.0 6.0 3.8 4.0 4.5 4.0 4.0 4.0 4.0 4.0 4.0 9.0 [NaCl] (mM) 450 450 300 50 50 50 1000 1000 550 50 50 50 50 50 250 50 Equil CVs Load CVs a 10 10 10 10 10 10 10 10 10 10 10 10 10 10 10 60 60a 60a 60a 60a 60a 20b 20b 20b 30a 30a 30a 30a 30a 20c 30a Elution CVs d 19 19d 19d 16e,f 16e,f 16e,f NAg NAg NAg 24e 24e 24e 24e 24e 13e 16e Strip CVs h 5h 5h 5h 5h 5h 10h 5h 10j 5h 5h 5h 5h 5h 5j 5h Gradient slope (mM CV−1 ) 55.3 55.3 63.2 59.4 59.4 59.4 NAg NAg NAg 60.4 60.4 60.4 60.4 60.4 57.7 59.4 a, load prepared by diluting glutathione affinity chromatography elution (GSH AC) product 3-fold into concentrated equilibration buffer; b load prepared by adjusting GSH AC product to desired conditions with small additions of M acetic acid and M NaCl stocks; c, prepared as in (a) with the addition of bovine serum albumin and λ-DNA spikes; d, [NaCl] at end of gradient was 1500 mM; e, [NaCl] at end of gradient was 10 0 mM; f, gradient followed by a CV step elution at 10 0 mM [NaCl]; g, Not applicable (NA); h, stripped with a pH 7.0, 100 mM Tris, 10 0 mM NaCl buffer; i, stripped with a pH 7.5, 100 mM Tris, 10 0 mM NaCl buffer; j, stripped with a pH 6.0, 50 mM citrate, 10 0 mM NaCl buffer nm The latter two wavelengths were used for pathlength correction purposes [22] The made measurements were employed to construct chromatographic traces These were used to design the pooling of the collected fractions and to identify fractions in need of further analysis Here, the fractions were pooled in a fashion yielding up to five pools containing flowthrough fractions (FT1– FT5), one pool containing wash fractions (W), and one pool containing strip fractions (S) The fractions collected during the elution of the RoboColumns were typically pooled in up to three different ways (i.e., E1–E3), unless stated otherwise Pools E1 and E2 contained the fractions collected in approximately the first and second half of the main elution peak, respectively Pool E3 contained all fractions included in pools E1 and E2 in addition to a few fractions collected after the last fraction included in pool E2 Pooling was carried out on a separate Tecan EVO 200 robotic station, operated by EVOware v2.8, which was equipped with an 8-channel disposable tip liquid handling arm Here, pools were generated at a desired volume by mixing equal volumes of fractions of interest, per RoboColumn, in separate wells of Thermo ScientificTM Armadillo PCR 96-well plates (Thermo Fisher Scientific Inc.) The fractions included in each pool are detailed in Table S1 (Supporting information) Generated pools and fractions were either analyzed immediately or stored at °C or –70 °C until their analysis Plates containing fractions and pools were sealed with Thermo ScientificTM NuncTM sealing tape (Thermo Fisher Scientific Inc.) For bind and elute chromatography, the RoboColumns were eluted in NaCl gradients with slopes of ∼55–63 mM CV−1 following their washing At the end of a gradient, the RoboColumns were stripped for CVs with a 100 mM Tris, pH 7.0, 10 0 mM NaCl buffer unless stated otherwise The same buffer was also used in flowthrough chromatography based separations to strip the RoboColumns at the end of their wash, with the exception of separation #9 which employed a 100 mM Tris, pH 7.5, 1500 mM NaCl buffer (Table 1) Linear elution salt gradients were simulated by multistep gradients wherein each step had a size of CV and a salt level (Csalt ) determined by the equation Csalt = Csalt,o + gradient slope × CVelution Here, Csalt,o is the salt level in the employed equilibration and wash buffers or the load, and CVelution corresponds to the number of CVs for which a RoboColumn was eluted for The steps in the gradient were generated by mixing low and high NaCl concentration buffers, per pH, in Axygen® 2.2 mL 96-well deep square well plates (Corning Life Sciences) at different ratios to obtain the desired Csalt Separation of full mature CVA21 virions from empty procapsids The full/empty CVA21 particle separation was tested in a range of mobile phase conditions for CEX resin PorosTM HS 50 (Thermo Fisher Scientific Inc.) Additional CEX resins CaptoTM S ImpAct, Capto SP ImpRes, Capto S, NuviaTM HR-S (Bio-Rad, CA, USA), and Nuvia S (Bio-Rad) were evaluated, along with the AEX resin Nuvia HP-Q (Bio-Rad) The CEX-based separations (i.e., #1–14) employed a 50 mM citrate buffer system at a pH range of 3.8–6.0, whereas the AEX-based separation (#16) employed a 50–70 mM Tris, pH 9.0 buffer system In both cases, the employed mobile phases included NaCl concentrations ([NaCl]) of 50 mM–1500 mM (Table 1) The equilibration, loading, and wash chromatography phases were carried out at pH, buffer, and [NaCl] conditions matching those of the equilibration buffer For separations #1–6, #10–14, and #16, the GSH AC product (feed) was diluted 3-fold in concentrated buffers to provide the load to the RoboColumns; this was loaded for 30–60 CVs (Table 1) The concentrated buffers were prepared at compositions (pH, buffer, and [NaCl]) matching those of the equilibration buffers post the 3-fold dilution of the GSH AC product For the AEX-based separation (#16), the Tris concentration was increased to 70 mM in the load compared to 50 mM in the equilibration buffer For separations #7–9, the GSH AC product was adjusted to match the equilibration buffer with the addition of small amounts of M citrate, pH 4.0 and M NaCl solutions it was loaded to the RoboColumns for 20 CVs (Table 1) Separation of full mature CVA21 virions from process related impurities The described CEX bind and elute RoboColumn methodology was also employed to perform column challenge experiments These were carried out by increasing the levels of impurities pre3 S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 sented to the resin and determining their impact on the separation of full mature virus particles from impurities, such as host cell DNA and bovine serum albumin (BSA) For this purpose, 600 μL Poros 50 HS RoboColumns were employed in separation #15 (Table 1) They were equilibrated for CVs before they were loaded for 20 CVs and washed for CVs with equilibration buffer The RoboColumns were then eluted for 13 CVs in a multi-step NaCl gradient with a slope of 57.7 mM CV−1 and stripped for CVs The mobile phases employed during the equilibration and wash of the RoboColumns were comprised of a 50 mM citrate, pH 4.0, 250 mM NaCl, 0.005% PS80 buffer system The strip employed a 50 mM citrate, pH 6.0, 10 0 mM NaCl, 0.0 05% PS80 buffer The generation of elution buffers took place as described in Section 2.3.1 to return 1/3 CV steps, each at an increasing NaCl level, while using a 50 mM citrate, pH 4.0, 10 0 mM NaCl, 0.0 05% PS80 buffer In these experiments, the GSH AC product was diluted 3-fold in concentrated buffers to match the equilibration buffer composition post dilution, as described in Section 2.3.1.1 The final material loaded to the RoboColumns was then generated by spiking the diluted GSH AC product with small volumes of concentrated BSA and λDNA (Thermo Fisher Scientific Inc.) stocks to final concentrations of 0.1 g L−1 and 200 ng mL−1 , respectively These corresponded to loading 1.2 mg BSA and 2.4 μg λ-DNA to the Poros 50 HS RoboColumns which represented a >100-fold increase of such impurities in a typical GSH AC product Absorbance measurements of the collected 200 μL fractions and their pooling took place as described in Section 2.3.1 small amounts of M citrate, pH 4.0 and M NaCl solutions The residence time across all steps was set to min, instead of used in HT scale 2.6 Sucrose density gradient centrifugation analysis of CVA21 process intermediates The presence of CVA21 empty procapsids and full mature virions in CEX chromatography loads and fractions from a large scale purification was verified via sucrose density gradient centrifugation performed Continuous sucrose gradients were prepared at 11 mL in Polyclear ultracentrifuge tubes (Seton Scientific, CA, USA) using 15 mM Tris, pH 8.0, 150 mM NaCl, 0.005% PS80 buffers containing sucrose at 15% (w/v) and 45% (w/v) Upon application of mL samples to the top of the tubes, the gradients were centrifuged at 360 0 rpm for 10 at 4°C using an OptimaTM -SE Ultracentrifuge (Beckman Coulter, CA, USA) Twelve fractions of equal volumes were then collected from the top of the gradients using a piston gradient fractionator (Biocomp Instruments, Canada) and stored at °C until further analysis 2.7 Analytical methods 2.7.1 Quantitative western blotting CVA21 full mature virion (VP4) and empty procapsid (VP0) contents in samples were determined via quantitative western blotting using a Sally SueTM system and a 12–230 kDa Sally SueTM Separation Module kit (Protein Simple, CA, USA) Here, it needs to be emphasized that while VP0 is included in both provirions and empty procapsids, the presence of the former in the purified CCCH samples is expected to be negligible [20] Therefore the VP0 measurements were indicative of the presence of empty propcapsids in tested samples Samples were prepared using an Anti–Rabbit Detection Module (Protein Simple) according to the manufacturer’s protocol and were denatured in a Mastercycler® Gradient (Eppendorf) for at 95 °C For their analysis, an anti–VP4 rabbit pAb (Lifetein LLC, NJ, USA), diluted to 20 μg mL−1 in Antibody Diluent (Protein Simple), was used Upon their preparation, the samples were loaded to the capillaries for sec, separated for 40 at 250 V, and immobilized for 250 sec This was followed by their exposure to antibody diluent for 23 min, to anti–VP4 rabbit primary antibody for 30 min, and to the anti–rabbit secondary antibody for 30 The capillaries were then imaged with the chemiluminescence detection settings and a s exposure time setting Peaks were integrated with a dropped lines method All samples were diluted with a concentrated Tris, pH 7.5 buffer, 0.005% PS80 to a final composition of ∼150 mM Tris, pH 7.5, 0.005% PS80 prior to their analysis Assay results were employed to determine yields via mass balancing 2.4 Stability of CVA21 in chromatography mobile phase conditions The impact of three factors on the stability of CVA21 was investigated: pH, [NaCl], and time For this purpose, GSH AC product was used, and it was diluted 3-fold in concentrated buffers to yield a final composition of 50 mM citrate buffers at 18 combinations of pH (3.8, 3.9, 4.0, 4.1, 4.2, 4.5) and [NaCl] (100 mM, 400 mM, 700 mM) conditions The starting GSH AC product was also included in this study as a control The 18 conditions and control were prepared in triplicate in separate wells of an Axygen 2.2 mL 96-well deep square well plate and upon their preparation the plate was sealed with Thermo Scientific Nunc sealing tape and shaken at 1100 rpm for 1.5 h at room temperature At the end of the incubation period, an aliquot was taken from each well of the plate and added to 0.5 mL MatrixTM 2D barcoded tubes (Thermo Fisher Scientific Inc.) which were stored at -70 °C until their analysis The plate was then sealed and left at room temperature for one day, under shaking, until a new aliquot was transferred to a second set of 0.5 mL Matrix 2D barcoded tubes, also stored at -70 °C until their analysis All used buffers contained 0.005% PS80 2.5 Large scale cation exchange column chromatography 2.7.2 Infectivity assay An automated, high-throughput viral imaging infectivity assay was used to measure CVA21 potency Briefly, in this assay, the tested samples and a CVA21 positive reference control were used to infect confluent 384 well tissue culture cell plates, which were planted with SK-MEL-28 cells (cat HTB-72; ATCC) Upon their infection and incubation, the plates were fixed, permeabilized, and stained with Hoechst 33342 (nuclei stain) (cat H3570; Thermo Fisher Scientific Inc.) Subsequently, the cells in the plates were immunostained with purified rabbit anti-CVA21 pAb (National Biologics Laboratory) and labeled with Alexa Fluor® 488 AffiniPure donkey anti-rabbit IgG (cat 711-545-152; Jackson ImmunoResearch Inc, PA, USA) The plates were then imaged for the stained nuclei and the fluorescently tagged viral protein on a BioTek CytationTM reader (Agilent Technologies, CA, USA) The images were analyzed on the reader’s software to count total (nuclei stain) and infected The Poros 50 HS CEX microscale purification method was scaled-up large scale using a 10 cm bed height column with a bed volume of 200 mL which was connected to an ÄKTA chromatography station, controlled by UNICORNTM v7 The column was first equilibrated for CVs using a 50 mM citrate, pH 4.0, 400 mM NaCl, 0.005% PS80 buffer, followed up by its loading for 27 CVs The column was then washed for CVs with a 25 mM citrate, pH 4.0, 50 mM NaCl, 0.0 05% PS80 buffer before it was eluted for CVs with a 25 mM citrate, pH 4.0, 800 mM NaCl, 0.005% PS80 buffer Finally, the column was stripped for CVs using a PBS buffer at pH 7.0, 10 0 mM NaCl and 0.005% PS80 Here, the load to the column was a process intermediate obtained by purifying clarified cell culture harvest with the GSH AC and intermediate IEX steps The intermediate IEX product was adjusted to match the composition of the equilibration buffer for the CEX step with the addition of S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 of pH 4.0, 540 mM NaCl on Capto SP ImpRes led to ∼100% and ∼13% flowthrough yields for full (Fig 1B) and empty (Fig 1E) particles, respectively Likewise, the same condition on Capto S ImpAct led to flowthrough yields of ∼100% and ∼0% for full (Fig 1A) and empty (Fig 1D) particles, respectively The aforementioned operating space became narrower with increased pH values suggesting that an optimal separation would need to employ mobile phases with a low pH While enteroviruses can be stable across a wide range of conditions [24], the employment of an acidic condition for separating full mature virus particles and empty procapsids, via CEX chromatography, led to concerns over potential infectivity losses for CVA21 These were addressed by the execution of a stability study at room temperature which evaluated the relationship between CVA21 infectivity and factors including liquid conditions (pH and NaCl concentration) and hold duration (two time points) (tagged viral protein) cells and these counts were used to calculate the percentage of infected cells in each well of a tested plate A dose response curve was then generated from the estimated percentage of infected cells for each sample and for the CVA21 reference standard in order to calculate the associated effective dose (ED) 50 Finally, the relative potency for each test sample was determined by taking the ratio between a sample’s ED50 to the reference control’s ED50 and reporting it as a percentage (%Response) 2.7.3 SDS-PAGE Samples were analyzed via gel electrophoresis using NuPAGETM 12% Bis-Tris 1.0 mm gels (Invitrogen, CA, USA) to track CVA21 empty procapsids and full mature virus particles (VP0 and VP2, respectively; VP4 could not be reliably tracked due to its molecular weight being close to the low limit of the gel) and proteinaceous impurities For this purpose, 700 μL of denaturing buffer was prepared by mixing 200 μL of NuPAGE Sample Reducing Agent (10X) (Invitrogen) and 500 μL of NuPAGE LDS Sample Buffer (4X) (Invitrogen) 14 μL and 26 μL of denaturing buffer and sample, respectively, were mixed in wells of a Thermo Scientific Armadillo PCR 96-well plate This was sealed with a Thermo Scientific Nunc sealing tape and centrifuged briefly at 30 0 rpm on a Sorvall Legend XTR centrifuge (Thermo Fisher Scientific Inc.) The PCR plate was then denatured in a Mastercycler Gradient (Eppendorf) for 10 at 70 °C Following denaturation, up to 25 μL of sample and μL of Mark12 Unstained Standard (Invitrogen) were loaded into separate lanes of a gel The prepared gels were electrophoresed for 50 at 200 V in a 1X MOPS running buffer, prepared from NuPAGE MOPS SDS Running Buffer (20X) (Invitrogen), and stained with a PierceTM Silver Stain Kit (Thermo Fisher Scientific Inc.) according to the manufacturer’s protocol, with a development time Gel images were generated on a Gel DocTM EZ System (Bio-Rad) with a Silver Stain autoexposure scan protocol VP0–VP4 were identified based on their expected molecular weight and annotated where possible by arrows 3.1.1 Impact of acidic conditions on CVA21 infectivity The performed stability study indicated an average decrease in CVA21 infectivity of 15.2% ± 10.6% between the two time points across the 18 liquid conditions tested (Fig 1G vs H) The pH and [NaCl] effects on CVA21 infectivity were investigated based on regression analysis (Table S2) and they were found to differ between the two time points; after the 1.5 h hold (Fig 1G), only NaCl concentration had a significant and positive effect, whereas after the 28 h hold (Fig 1H), both pH and NaCl concentration had positive and almost equal effects on CVA21 infectivity Moreover, the relationship between infectivity and these two factors was stronger at the second time point compared to the first one (i.e., %R2 of ∼17% and ∼49% at the first and second time points, respectively, in Table S2) This implied that pH and [NaCl] affected CVA21 infectivity more prominently at increased hold times at room temperature The loss of infectivity as a function of time was also observed in the GSH AC product control sample which was buffered at pH 8.0 (Fig 1G and H) The employment of one-way analysis of variance to compare between the measured infectivities of the control sample and of the tested 18 samples, at the first time point, showed no significant difference between the 19 samples (Table S3) Hence, the time dependent infectivity losses in Fig 1G and H were not specific to the tested acidic conditions alone; instead they also included inherent, short term infectivity losses for CVA21 at room temperature Based on these results, the application of CEX chromatography to purify full mature virions form empty procapsids at acidic conditions was deemed to be a viable approach for CVA21 purification since no significant infectivity losses are expected over the short duration of the CEX step (∼5 h at large scale) 2.7.4 Total protein, DNA, and bovine serum albumin analytics Quant-iTTM PicoGreenTM dsDNA (Invitrogen, CA, USA) and PierceTM Coomassie Plus (Bradford) (Thermo Fisher Scientific Inc.) assays were deployed as per the manufacturer’s instructions BSA quantitative western blotting analysis was performed as described in [23] Results and discussion 3.1 Identification of cation exchange chromatography for separation of CVA21 full mature virions and empty procapsids 3.2 Poros 50 HS chromatography for separation of CVA21 full mature virions and empty procapsids RoboColumn resin screening of GSH AC product from an early static culture virus production process supported the application of cation exchange chromatography for separating CVA21 empty procapsids from full mature virions, as opposed to anion exchange and hydrophobic interaction chromatography (data not shown) These early results were further corroborated via screening HT batch chromatography experiments (Fig 1A–F) Full mature virions bound to cation exchangers Capto S ImpAct and SP ImpRes only at pH 4.0, 420 mM NaCl (Fig 1A and B, respectively), whereas binding of empty procapsids to these resins was stronger across a wider range of tested conditions (Fig 1D and E, respectively) Conversely, the multimodal resin Capto MMC ImpRes bound more strongly both types of particles across a wider range of test conditions (Fig 1C and F) The binding differences between the two particle types outlined a pH and NaCl concentration operating space resulting in nearly complete separation of CVA21 full mature virions from empty procapsids on each resin For example, employing a binding condition The early chromatography screening and stability testing results were followed by the characterization of the CEX-based CVA21 purification using RoboColumns Here, Poros 50 HS resin was employed in bind and elute mode (separations #1–6 in Table 1) due to its large pore size characteristics [25], rendering it better suited to adsorb large solutes, such as CVA21 particles The chromatograms from separations #1–6 (Figs 2A and S1) demonstrated the excellent repeatability of the RoboColumn technique and yielded valuable information At pH 5.0 and 6.0 (separations #5 and #6, respectively), only a single peak was observed in the elution gradient, whereas at pH 3.8–4.5 (separations #1–4, respectively) two peaks were observed; one in the gradient and one in the strip For each separation, the fractions in each peak were pooled into elution (E3) and strip (S) pools (Table S1) The areas of the elution and strip peaks in the chromatograms increased and decreased, respectively, with increasing pH This indicated the presence of two solute populations with their retention S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Preliminary high throughput cation exchange batch chromatography screening results for Coxsackievirus A21 and its infectivity dependence on liquid conditions and time: (A)–(C) Flowthrough yields for full mature virions, as a function of binding pH and NaCl concentration ([NaCl]), for resins Capto S ImpAct, Capto SP ImpRes, Capto MMC ImpRes, respectively; (D)–(F) Flowthrough yields for empty procapsids, as a function of binding pH and NaCl concentration ([NaCl]), for resins Capto S ImpAct, Capto SP ImpRes, Capto MMC ImpRes, respectively; (G) and (H) %Response, depicting CVA21 infectivity based on the deployed viral imaging infectivity assay, as a function of pH for a 1.5 h and 28 h hold, respectively, at room temperature In (A)–(F) the yields (z-axis) are averages of duplicates The colorbar in (C) denotes the color scale across (A)–(C) The colorbar in (F) denotes the color scale across (D)–(F) In (G) and (H), symbols (o), ( ), and (♦) correspond to NaCl concentrations of 100 mM, 400 mM, and 700 mM, respectively, and symbol ( ) corresponds to a non-acidic control sample Error bars correspond to ±1 standard deviation (sd) being strongly affected by pH; a weaker binding one, eluting in the salt gradient, and a stronger binding one, eluting in the strip The composition of the two populations was determined via SDSPAGE analysis of pools E3 and S (Fig 3A) Here, it is noted that this analysis did not employ concentration normalizations during gel loading and hence its results also reflected volumetric concentrations as depicted by the generation of pools from the collected fractions (Table S1) Furthermore, silver stain also stains single and double stranded DNA and RNA (e.g., [26]) and for samples which were rich in full mature virions this resulted to the observation of a band at the top of the loaded gel lanes which was attributed to genomic RNA of CVA21 For separations #1–4, E3 contained primarily full mature CVA21 particles (abundant VP2 band and little to no presence of VP0 band), whereas S contained empty procapsids and small amounts of full mature virions (abundant VP0 band and presence of VP2 band) (Fig 3A) In contrast, for separations #5–6, E3 contained both full mature CVA21 particles and empty procapsids, and S contained neither of the two (Fig 3A) The composition of these two populations was also investigated by quantitative western blotting analyses of the FT, W, E3 and S pools (Fig 2B) For all six separations, the FT and W pools contained ∼0% of the full and empty particles included in the loaded GSH AC product Hence, all binding conditions, spanning a pH range of 3.8–6.0, resulted in high binding of CVA21 particles to Poros 50 HS Elution yields, as depicted by the E3 pool yields, varied between ∼75%–∼100% and increased with increasing pH Strip yields, as depicted by the S pool yields, decreased with increasing pH (Fig 2B) These led to mass balance closures well in excess of 80% for full mature CVA21 virions Hence, the elution yields of CVA21 full particles were high and varied within a narrow range However, for separations #1–6, the E3 pool yields for the empty procapsids varied between ∼0% and 60% and increased rapidly with increasing pH (Fig 2B) Mass balance closures for these particles (∼30%–∼70%) were poorer than those observed for the full CVA21 particles and the unaccounted empty particles were considered to be irreversibly bound to the resin While this would have a negative effect on the re-use of a column, unless the empty procapsids were removed through a cleaning in place strategy, here their irreversible binding was considered to be a desirable feature of this step Hence, a separation between CVA21 full mature virions and empty procapsids was also achieved using packed bed column chromatography, and its resolution depended on pH This agreed with the early batch-based chromatography HT screens (Fig 1A–F) The results generated from separations #1–6 provided sufficient information to derive the conclusion that attempting to purify full mature CVA21 particles from empty particles at less acidic pH conditions led to their co-elution in the salt gradient This was primarily due to pH effects on the retention of the empty procap- S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Bind and elute Poros 50 HS high throughput RoboColumn chromatography results for the separation of Coxsackievirus A21 full mature virions from empty procapsids: (A) 3D plot of chromatographic traces of recorded absorbance at 260 nm as a fraction number (y-axis) from six separations (#1–6), each at a different pH (x-axis) Lines (-) and (—) denote absorbances (Abs.) and salt levels, respectively, normalized by their maximum, and symbols (◦) and ( ) denote duplicated experiments (R1 and R2) The z-axis is a normalized scale from to where denotes the maximum; (B) Bar plot of elution and strip yields and mass balances for full mature virions and empty procapsids (each bar corresponds to a different pH/separation) Error bars correspond to ±1 standard deviation (sd); (C) Elution salt levels of main elution peak as a function of pH The salt level was determined by identifying the fraction associated to the beginning of the elution peak sids which decreased more as a function of the pH compared to the retention of the full particles This suggested that the encapsidation of genome in the full particles led to an increase in their negative net charge which led to electrostatic repulsions with CEX resins and reduced retention compared to the empty particles This agrees with earlier studies for a selection of picornaviruses, such as enterovirus 71, which observed the presence of fewer negatively charged surface patches for empty particles compared to full particles [27] The difference in the retention between the two CVA21 particle types on Poros 50 HS was exploited to define operating windows for establishing the polishing purification step for CVA21 and to deploy it at large scale sufficiently high to enable the selection of conservative NaCl concentrations in binding conditions (e.g., 200 mM–650 mM) to avoid any loss of CVA21 particles in the flowthrough, while allowing for the robust preparation of mobile phases and ease of implementation The latter is particularly important when taking into consideration that at large scale processing the product of the GSH AC step, eluted in a 100 mM NaCl buffer, was loaded directly to the following IEX step, which was run in flowthrough mode Hence, adopting a pH binding condition within 3.8–5.0 for the CEX polishing step would only require the pH adjustment of the intermediate product and not its dilution (typically undesired at large scale processing) While the Poros 50 HS operating pH range of 3.8–5.0 led to optimal and robust conditions for binding of GSH AC purified CVA21 particles, the optimal elution pH range for separating full mature virions from empty procapsids was narrower Pool E3 contained quantifiable amounts of empty procapsids (Fig 2B) at a pH between 4.2 and 5.0 (∼16%–∼60%) Conversely, elution pH values between 3.8 ≤ pH < 4.2 led to the complete separation of CVA21 full particles from empty particles (Figs 2B and 3A) This supported the selection of these pH values as the optimal elution pH range Such elution pH conditions had additional benefits that rendered them well suited to large scale processing SDS-PAGE analysis for separations #1 (Fig 3B) and #2 (Fig 3C) supported that within this elution pH range, full mature CVA21 virions could be eluted in a concentrated form, and collected with robust collection windows, via the application of a single step elution method with a step at high salt level Finally, the near optimal separation of CVA21 mature virions from empty procapsids at a pH range of 4.2–4.5 could also render this pH range as a viable alternative for eluting CVA21 At such 3.2.1 Operating windows for separating CVA21 full mature virions and empty procapsids via bind and elute chromatography Retention trends were generated (Fig S2) using the bind and elute chromatograms in Fig 2A to describe the interaction between the GSH AC purified CVA21 particles and the Poros 50 HS resin A quadratic relationship was, therefore, derived describing the dependence of elution salt on pH for the main elution peak (Fig 2C) and binding at a salt level below the fitted line would result to the binding of CVA21 particles, in the GSH AC product, to the Poros 50 HS resin Here, it needs be emphasized that in Fig 2C the salt levels represent their concentration in elution buffers at the inlet of a column Hence, a safety factor of at least ∼60 mM NaCl (i.e., approximately equal to the employed gradient slope in mM CV−1 ) would need be considered when choosing the binding NaCl concentration based on these results At a pH range of 3.8–5.0, the required elution salt levels varied between ∼350 mM and ∼800 mM NaCl (Fig 2C) These were S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Bind and elute Poros 50 HS high throughput RoboColumn chromatography SDS-PAGE results for the separation of Coxsackievirus A21 full mature virions from empty procapsids: (A) Gel images of GSH affinity chromatography product (feed), 3-fold diluted feed in concentrated equilibration buffer (load), elution pool (E3), and strip pool (S) for six separations (#1–6), each at a different pH; (B)–(D) Gel images of fractions comprising pool E3 for six separations (#1–6), each at a different pH, respectively In (A) and (B) text above each lane denotes the identity of the tested sample pH conditions, pool E3 corresponded to empty procapsid elution yields of ∼16% and ∼35%, respectively (Fig 2B) However, for these two conditions, the E3 pools were comprised of fractions 68–84 and 71–81, respectively (Table S1), which included fractions at the tail of the corresponding elution peaks (Figs 2A and S1) The late eluting fractions contained decreasing and increasing amounts of full mature virions and empty procapsids, respectively (Fig 3D and E as indicated by the intensity of the VP2 and VP0 bands) Excluding a few fractions from the tail of the elution peaks (e.g., fractions 76, 77 and 79–81 for separations #3 and #4, respectively) would result in product pools free of empty procapsids, at the cost of a marginal reduction in the elution yields for full CVA21 particles; the majority of the full particles eluted in a small number of fractions at lower salt levels than the empty procapsids (Fig 3D and S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Flowthrough mode Poros 50 HS high throughput RoboColumn chromatography SDS-PAGE results for the separation of Coxsackievirus A21 full mature virions from empty procapsids: (A) Gel images of 3-fold diluted GSH affinity chromatography product adjusted to the desired pH and NaCl concentration conditions (load), flowthrough pool (FT), wash pool (W), and strip pool (S) for three separations (#7–9, respectively), each at a different binding condition (pH and NaCl concentration); (B) Gel images of load and 13 fractions collected during the loading of the column for separation #9 employing a binding condition of pH 4.5 and 550 mM NaCl In (A) and (B) text above each lane denotes the identity of the tested sample E) Hence, purifying CVA21 full particles at a range of 4.2 ≤ pH ≤ 4.5 would be near optimal albeit with the requirement of more stringent peak collection control to avoid co-purifying empty procapsids in the CEX CVA21 product pool bind and step elution However, the latter offers the significant advantage of product concentration via volumetric reduction, which is desirable for subsequent downstream unit operations This contributed to the selection of bind and elute Poros 50 HS chromatography for the polishing of CVA21 at large scale 3.2.2 Separation of CVA21 full mature virions from empty procapsids via flowthrough chromatography The stronger binding of CVA21 empty procapsids to the Poros 50 HS resin at acidic conditions, compared to its full mature virions, made possible the purification of the GSH AC product via flowthrough mode CEX chromatography Separations #7–9 (Table 1) were carried out at pH conditions of 3.8, 4.0 and 4.5, respectively (Fig S3) Here, the employed salt level was determined based on the retention trends elucidated from the bind and elute experiments (Fig 2C) and the SDS-PAGE analysis results in Fig 3B, C and E High flowthrough yields for full mature CVA21 particles were obtained for separations #7–9 based on quantitative western blotting analysis (i.e., 92.7% ± 2.9%, 91.7% ± 4.1% and 84.8% ± 8.0%, respectively) The highly robust nature of CVA21 full and empty particle separation at pH ≤ 4.0 was also supported by separations #7 and #8; at this pH range, the tested flowthrough pools were free of empty procapsids (Fig 4A) Separation #9, carried out at a pH of 4.5, led to the inclusion of a small amount of empty procapsids in the flowthrough product pool (13.6%±7.3% and Fig 4A) This corresponded to a ∼3-fold reduction in the co-purified empty procapsids in the E3 pool of the bind and elute separation #4 (Figs 2B and 3A) The breakthrough of empty procapsids for separation #9 was tracked via SDS-PAGE analysis (Fig 4B), which showed that empty procapsids were flowing through at low amounts during early stages of column loading and their abundance increased with increasing loading Hence, their further reduction would require fine-tuning of both pH and [NaCl] in the binding conditions instead of modulating the amount of the loaded GSH AC product alone Despite this, those pH conditions that were found to be optimal in a bind and elute based purification (i.e., pH < 4.2) were also optimal when deployed in a flowthrough mode based purification Purifying CVA21 GSH AC product with Poros 50 HS, with optimal bind and elute or flowthrough mode chromatography conditions, led to product pools with high full mature CVA21 particle yields and free of empty procapsids Hence, both purification modes were viable Considering large scale unit operations, implementing a polishing step in flowthrough mode is easier than in 3.3 Large scale CVA21 full mature virion purification via Poros 50 HS bind and elute chromatography The polishing of CVA21 via Poros 50 HS bind and elute chromatography, at pH of 4.0, was verified at large scale using a 200 mL column (Fig 5A) The large scale purification process deployed the CEX polishing step after the preceding IEX and GSH AC steps Bound CVA21 particles were step-eluted from the Poros 50 HS column at 800 mM NaCl This led to the observation of a single peak containing concentrated full mature CVA21 particles and no empty particles The latter were recovered during the stripping of the column with a neutral pH and high salt buffer This behavior, along with the absence of any particles in the collected flowthrough (Fig 5B), agreed with the observation made from the HT scale experiments (Figs 2B and 3A, C) Good agreement was also observed across scales for full mature virion yields with large scale yields of ∼91% and ∼84% yields, based on quantitative western blotting and infectivity assays, respectively The high product volumes generated from the large scale run enabled the use of sucrose density gradient centrifugation analysis to verify the CEX-based separation of full CVA21 particles from empty particles The process intermediate, which was loaded to the 200 mL Poros 50 HS column, was shown to contain both particle types (Fig 6A), whereas the Poros 50 HS elution product was free of empty particles (Fig 6B) These results demonstrated further the scalability of the HT scale results and provided additional confirmation for the performance of the selected conditions; they led to high elution yields and a complete separation of full CVA21 particles from empty ones in a robust and easy to implement purification at large scale 3.4 Purification of CVA21 full mature virions from process related impurities via Poros 50 HS bind and elute chromatography Apart from separating full mature virions from empty procapsids, the CEX step was also determined to flow through small S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Bind and elute Poros 50 HS large scale chromatography results for the separation of Coxsackievirus A21 full mature virions from empty procapsids: (A) Chromatographic trace at 280 nm on the left-hand side y-axis and conductivity, pH traces on the right-hand side y-axis The x-axis represents column volumes; (B) SDS-PAGE analysis of fractions collected across the entire loading phase (FT), elution phase (E) and strip phase (S) of the chromatogram in (A) Text above each lane denotes the identity of the tested sample Fig SDS-PAGE analysis of fractions collected during sucrose density gradient centrifugation for: (A) Starting material (CEX Load) purified by the large scale bind and elute Poros 50 HS polishing step; (B) Elution product pool (CEX Elution) from the polishing step In both (A) and (B) the second lane shows the sample that was analyzed by sucrose density gradient centrifugation and lanes B1–B12 show the fractions collected during their sucrose density gradient centrifugation analysis from the top to the bottom collected layers In (A) and (B), text above each lane denotes the identity of the tested sample Boxes with dashed lines denote the location of empty procapsids and full mature virions in the collected fractions amounts of persistent high molecular weight proteinaceous impurities, which were not removed by the preceding GSH AC and intermediate IEX steps This was observed while purifying CCCH from upstream process A and testing the generated products with overexposed SDS-PAGE gels (Fig S4) This trend was also observed when purifying CCCH from upstream process B; for separation #9 (Fig 4B) the collected 20 CV flowthrough pool contained both CVA21 particles and faint bands of protein impurities at similar molecular weights to those observed in Fig S4 This supported further the operation of the Poros 50 HS step in bind and elute mode rather than flowthrough mode for CVA21 purification The observation that the bind and elute CEX step contributed to a further reduction of process related impurities led to the execution of studies aiming to challenge its purification potential Here, large amounts of BSA and λ-DNA were spiked to GSH AC product, as described in Section 2.3.1.2, and the CEX purification was performed with a binding condition of pH 4.0, 250 mM NaCl and a NaCl gradient of 58 mM CV−1 BSA represented a major process related impurity to be removed by the downstream process, present due to the inclusion of bovine calf serum in the cell culture, whereas λ-DNA represented a molecularly-distinct type of contaminant (i.e., host cell DNA) in need of removal from the final purified product It needs be highlighted that the employed GSH AC product purified in this study was generated by upstream process A and as shown in Fig S4 it included a low amount of empty procapsids The overlay of chromatographic traces from the HT total protein and DNA assays led to the observation of three peaks in the salt gradient and two peaks in the neutral pH column strip (Fig 7) Their pooling (Fig 7) was followed by analytical testing to determine the presence of full CVA21 particles and BSA via quantitative western blotting The former were observed only in elution pool E2, leading to an elution yield of 96.7% ± 5.2% BSA was observed in pools E2 and E4 with increased presence at higher salt levels (i.e., 17.6 μg ± 1.0 μg and 370.0 μg ± 29.3 μg, respectively, based on quantitative western blotting analysis) Pool E1 was not tested 10 S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig High throughput RoboColumn chromatography results from challenging the bind and elute Poros 50 HS step by applying it to the purification of GSH affinity chromatography Coxsackievirus A21 product spiked with large amounts of BSA and λ-DNA (separation #15) The left hand-side y-axis shows overlaid concentration traces (-) plotted against the number of collected fractions as generated by the Bradford (Protein, (◦)) and PicoGreen (dsDNA, ( )) assays The right hand-side y-axis shows the NaCl concentration (-) during the separation Insert at the top left hand-side corner shows a zoomed-out view of the entire separation to focus on the column load and wash phases Double headed arrows (↔) denote the beginning and end points of flowthrough (FT), wash (W), elution (E1–E4) and strip (S) pools, also separated by dashed vertical lines Reported concentrations are averages of two replicates only for 2 log increase of protein content in a typical GSH AC product (Section 2.3.1.2) Similarly, due to the absence of additional solutes that could elute in the strip and also return a strong HT DNA assay signal, the DNA rich peak in the strip (Fig 7) was attributed to λ-DNA Strong binding of DNA to Poros 50 HS has been reported previously [28] without an explanation of the underlying binding mechanism The HT DNA assay trace also displayed a small peak for fractions coinciding with pool E2, suggesting the co-elution of DNA and CVA21 full particles (Fig 7) However, earlier studies indicated that the HT DNA assay could track the presence of full mature CVA21 particles during the CEX polishing step and the generated signal in pool E2 was therefore inclusive of the eluting CVA21 particles Despite this, the HT DNA assay signal in pool E2 accounted 3.5 CVA21 full mature virus particle purification with alternative CEX media Although early batch chromatography screening results demonstrated that multiple CEX stationary phases could lead to a good separation between full and empty CVA21 particles (Fig 1A–F), these experiments employed starting material from an early virus production process In order to confirm this important observation GSH AC product from virus production process B was employed to assess five alternative cation exchangers in separations #10–14 (Table 1), carried out at a pH of 4.0 Similar to the Poros 50 HS bind and elute separation at a pH 4.0 (separation #2 in Table 1), these separations resulted in one peak in the salt gradient and a second one in the strip (Figs 8A and S5) This behavior resembled the one observed in separations #1–6 (Figs 2A 11 S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 Fig Bind and elute high throughput RoboColumn chromatography results for the separation of Coxsackievirus A21 full mature virions from empty procapsids with resins Capto S ImpAct, Capto SP ImpRes, Capto S, Nuvia HR-S, Nuvia S, and Nuvia HP-Q (separations #10–14 and #16, respectively): (A) 3D plot of chromatographic traces of recorded absorbance at 260 nm as of fraction number (y-axis) from six separations, each with a different resin (x-axis) Lines (-) and (—) denote absorbances (Abs.) and salt levels, respectively, normalized by their maximum, and symbols (◦) and ( ) denote duplicated experiments (R1 and R2) The z-axis is a normalized scale from to where denotes the maximum; (B) Bar plot of elution and strip yields and mass balances for full mature virions and empty procapsids (each bar corresponds to a different resin/separation) Error bars correspond to ±1 standard deviation (sd); (C) SDS-PAGE gel images of GSH affinity chromatography product (feed), 3-fold diluted feed in concentrated equilibration buffer (load), elution pool (E3), and strip pool (S) for six separations, each at a different resin In (C), the gel image of Nuvia HP-Q shows the load and the fractions comprising of the main elution peak for this separation (#16) instead of the feed, E3 and S In (C) text above each lane denotes the identity of the tested sample and S1) For separations #10–14, the fractions in the observed two peaks were combined in elution and strip pools, which were then analyzed via quantitative western blotting (Fig 8B) and SDS-PAGE (Fig 8C) These showed that all alternative cation exchange resins led to high elution yields for the full mature CVA21 particles and to elution product pools virtually free of empty procapsids as they eluted in the strip The estimated elution yields were higher for the stationary phases with bead sizes similar to Poros 50 HS (∼50 μm) (i.e., Capto S ImpAct, Capto SP ImpRes, and Nuvia HR-S led to >90% yields), compared to those with bead sizes of ∼90 μm (i.e., Capto S and Nuvia S led to ∼70% yields) The retention trends from these separations (Fig S6) showed that the latter group of resins eluted the full mature CVA21 particles at lower salt levels compared to the former group which eluted them at similar NaCl concentrations to Poros 50 HS Consequently, the lower elution yields for the full CVA21 particles for resins Capto S and Nuvia S could not be attributed solely to their strong electrostatic binding to these resins since this would have resulted in elution of the full particles at higher salt levels compared to resins Capto S ImpAct, Capto SP ImpRes and Nuvia HR-S The elution behavior of this type of particle was therefore most likely affected by multiple factors in addition to a resin’s bead size Full mature CVA21 particle mass balances for separations #10–14, based on the elution and strip pool yields alone, were in excess of 95% (Fig 8B) Hence, similarly to the Poros 50 HS purification (separation #2), the tested alternative cation exchangers bound nearly all full mature CVA21 particles in the loaded GSH AC product at a pH of 4.0 The complete binding of the CVA21 particles and their separation into empty and full particle populations, for five different cation exchangers, suggested that the cation exchange modality itself was the main driving force enabling the highly efficient and effective purification of CVA21 particles, more so than the intrinsic properties of the different stationary phases This enables its wide applicability, with respect to the applied chromatography media, and could extend it to include membrane absorbers and monoliths, in addition to the diffusive media tested here On the contrary, AEX chromatography was determined to have poor applicability for the purification of CVA21 particles The AEX resin Nuvia HP-Q, run at pH 9.0 in bind and elute mode (separation #16 in Table 1), returned high elution yields for full particles (Fig 8B), which co-eluted, however, with the empty particles, also shown via SDS-PAGE analysis (Fig 8C) Furthermore, the binding of CVA21 particles to AEX resins decreased significantly as the pH decreased to 8.0, even at low binding salt levels, indicating that binding could occur only within a narrow range of conditions (data not shown) This rendered AEX chromatography as potentially unfavorable for the separation of full CVA21 particles from empty The purification of GSH AC product, with the additional CEX resins at a pH of 4.0, was determined to be very similar to the Poros 50 HS based purification; all CEX resins returned high binding of CVA21 particles, which were separated into full mature virions and empty procapsids The former eluted in the salt gradient, with high yields, whereas the latter eluted at a high pH and high salt concentration buffer The difference in the retention of the two 12 S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 particle types would potentially enable the deployment of these resins in a similar fashion to Poros 50 HS at large scale (i.e., step elution at a high salt level) These observations supported the wide applicability of CEX for purifying CVA21 and had a direct impact on the large scale CVA21 purification process since the identification of multiple resins can mitigate process disruptions due to, for example, procurement challenges Murphy R Poplyk: Methodology, Investigation, Data curation, Visualization, Writing – original draft Andrew R Swartz: Methodology, Validation, Resources, Investigation, Data curation, Visualization, Writing – review & editing Richard R Rustandi: Methodology, Investigation, Data curation, Writing – review & editing Rachel Thompson: Methodology, Investigation, Data curation, Writing – original draft Sheng-Ching Wang: Resources, Project administration Conclusions Acknowledgments Empty viral particles represent a challenging product related impurity in viral vaccines and gene therapy drug substance production Their separation from full mature particles typically requires techniques that are not amenable to large scale processing While anion exchange chromatography based purifications can be used, these are often limited by a low selectivity between the two particle types and a significant purity and yield trade-off due to peak cutting The application of anion exchange chromatography was shown to be ineffective in separating full mature CVA21 virions from empty procapsids Conversely, cation exchange chromatography, applied in either bind and elute or flowthrough mode, led to high yields of full mature CVA21 particles which were free of empty procapsids In bind and elute mode, the CEX step was able to be readily deployed at large scale, due to its robustness, ease of implementation, and enabling efficient subsequent processing While the main outcome of this step was separation of empty CVA21 procapsids from full mature particles, it was also shown that the CEX step contributed to the robustness of the downstream purification process; it removed proteinaceous impurities, which were not entirely cleared by previous purification steps, and could separate the full mature particles from prominent process related impurities in the presence of feed-stream variability These observations supported the establishment of cation exchange chromatography as a polishing step in the purification of CVA21 Furthermore, preliminary evidence that this effective purification modality extended to multiple cation exchange resins and alternate enteroviruses, support its applicability as a general technique for enterovirus-based therapeutic products Supporting Information: Additional supporting information may be found in the online version of this article This includes Tables S1–S3 and Figs S1–S6 The authors acknowledge Shieh Yvonne for carrying out the sucrose density gradient centrifugation analysis, Jimmy Devlin for carrying out the infectivity analysis, Michael A Winters, Joseph G Joyce and Rebecca A Chmielowski for their assistance in preparing this manuscript Supplementary materials Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.chroma.2022.463259 References [1] J Baggen, H.J Thibaut, J.R Strating, F.J van Kuppeveld, The life cycle of non-polio enteroviruses and how to target it, Nat Rev Microbiol 16 (6) (2018) 368–381 [2] P.M Chumakov, V.V Morozova, I.V Babkin, I.K Baikov, S.V Netesov, N.V Tikunova, Oncolytic enteroviruses, Mol Biol 46 (5) (2012) 639–650 [3] H Fukuhara, Y Ino, T Todo, Oncolytic virus therapy: a new era of cancer treatment at dawn, Cancer Sci 107 (10) (2016) 1373–1379 [4] D Masemann, Y Boergeling, S Ludwig, Employing RNA viruses to fight cancer: novel insights into oncolytic virotherapy, Biol Chem 398 (8) (2017) 891–909 [5] N.E Annels, D Mansfield, M Arif, C Ballesteros-Merino, G.R Simpson, M Denyer, S.S Sandhu, A.A Melcher, K.J Harrington, B Davies, G Au, M Grose, I Bagwan, B Fox, R Vile, H Mostafid, D Shafren, H.S Pandha, Phase I trial of an ICAM-1-targeted immunotherapeutic-coxsackievirus A21 (CVA21) as an oncolytic agent against non muscle-invasive bladder cancer, Clin Cancer Res 25 (19) (2019) 5818–5831 [6] S Bradley, A.D Jakes, K Harrington, H Pandha, A Melcher, F Errington-Mais, Applications of coxsackievirus A21 in oncology, Oncolytic Virotherapy (2014) 47 [7] N Jayawardena, J.T Poirier, L.N Burga, M Bostina, Virus–receptor interactions and virus neutralization: insights for oncolytic virus development, Oncolytic Virotherapy (2020) [8] P Jiang, Y Liu, H.C Ma, A.V Paul, E Wimmer, Picornavirus morphogenesis, Microbiol Mol Biol Rev 78 (3) (2014) 418–437 [9] A Mullard, Gene therapy community grapples with toxicity issues, as pipeline matures, Nat Rev Drug Discov (2021) [10] M.M Hankaniemi, O.H Laitinen, V.M Stone, A Sioofy-Khojine, J.A Määttä, P.G Larsson, V Värjomaki, H Hyöty, M Flodström-Tullberg, V.P Hytönen, Optimized production and purification of Coxsackievirus B1 vaccine and its preclinical evaluation in a mouse model, Vaccine 35 (30) (2017) 3718–3725 [11] C.C Liu, M.S Guo, F.H.Y Lin, K.N Hsiao, K.H.W Chang, A.H Chou, Y.C Wang, Y.C Chen, C.S Yang, P.C.S Chong, Purification and characterization of enterovirus 71 viral particles produced from vero cells grown in a serum-free microcarrier bioreactor system, PLoS One (5) (2011) e20 05 [12] M.G Moleirinho, R.J Silva, P.M Alves, M.J Carrondo, C Peixoto, Current challenges in biotherapeutic particles manufacturing, Expert Opin Biol Ther 20 (5) (2020) 451–465 [13] M.M Segura, A.A Kamen, A Garnier, Overview of current scalable methods for purification of viral vectors, Viral Vectors Gene Ther (2011) 89–116 [14] W Qu, M Wang, Y Wu, R Xu, Scalable downstream strategies for purification of recombinant adeno-associated virus vectors in light of the properties, Curr Pharm Biotechnol 16 (8) (2015) 684–695 [15] T Okada, M Nonaka-Sarukawa, R Uchibori, K Kinoshita, H Hayashita-Kinoh, Y Nitahara-Kasahara, S Takeda, K Ozawa, Scalable purification of adeno-associated virus serotype (AAV1) and AAV8 vectors, using dual ion-exchange adsorptive membranes, Hum Gene Ther 20 (9) (2009) 1013–1021 [16] M Urabe, K.Q Xin, Y Obara, T Nakakura, H Mizukami, A Kume, K Okuda, K Ozawa, Removal of empty capsids from type adeno-associated virus vector stocks by anion-exchange chromatography potentiates transgene expression, Mol Ther 13 (4) (2006) 823–828 [17] M Hebben, Downstream bioprocessing of AAV vectors: industrial challenges & regulatory requirements, Cell Gene Ther Insights (2018) 131–146 [18] A.R.K Venkatachalam, M Szyporta, T.K Kiener, P Balraj, J Kwang, Concentration and purification of enterovirus 71 using a weak anion-exchange monolithic column, Virol J 11 (1) (2014) 1–8 Declaration of Competing Interest Spyridon Konstantinidis reports a relationship with Merck & Co Inc that includes: employment and equity or stocks Murphy R Poplyk reports a relationship with Merck & Co Inc that includes: employment and equity or stocks Andrew R Swartz reports a relationship with Merck & Co Inc that includes: employment and equity or stocks Richard R Rustandi reports a relationship with Merck & Co Inc that includes: employment and equity or stocks Rachel Thompson reports a relationship with Merck & Co Inc that includes: employment and equity or stocks ShengChing Wang reports a relationship with Merck & Co Inc that includes: employment and equity or stocks Spyridon Konstantinidis has patent #US20210187049A1 pending to Merck Sharp and Dohme Corp Murphy R Poplyk has patent #US20210187049A1 pending to Merck Sharp and Dohme Corp Andrew R Swartz has patent #US20210187049A1 pending to Merck Sharp and Dohme Corp CRediT authorship contribution statement Spyridon Konstantinidis: Conceptualization, Methodology, Software, Formal analysis, Investigation, Data curation, Visualization, Writing – original draft, Writing – review & editing 13 S Konstantinidis, M.R Poplyk, A.R Swartz et al Journal of Chromatography A 1676 (2022) 463259 [19] J.Y Chang, C.P Chang, H.H.P Tsai, C.D Lee, W.C Lian, I.H Sai, C.C Liu, A.H Chou, Y.J Lu, C.Y Chen, P.H Lee, J.R Chiang, P.C.S Chong, Selection and characterization of vaccine strain for enterovirus 71 vaccine development, Vaccine 30 (4) (2012) 703–711 [20] E.M Curtis, S Konstantinidis, M Poplyk, A.R Swartz, M.D Wenger, inventors; Merck Sharp and Dohme Corp, assignee Purified compositions of enteroviruses and methods of purification with glutathione affinity chromatography US patent 17/125002 2020 Dec 17 [21] S Konstantinidis, H.Y Goh, J.M Martin Bufájer, P de Galbert, M Parau, A Velayudhan, Flexible and accessible automated operation of miniature chromatography columns on a liquid handling station, Biotechnol J 13 (3) (2018) e1700390 Mar [22] E.L McGown, D.G Hafeman, Multichannel pipettor performance verified by measuring pathlength of reagent dispensed into a microplate, Anal Biochem 258 (1) (1998) 155–157 [23] J.W Loughney, C Lancaster, S Ha, R.R Rustandi, Residual bovine serum albumin (BSA) quantitation in vaccines using automated Capillary Western technology, Anal Biochem 461 (2014) 49–56 [24] R.J Salo, D.O Cliver, Effect of acid pH, salts, and temperature on the infectivity and physical integrity of enteroviruses, Arch Virol 52 (4) (1976) 269–282 [25] Y Wu, J Simons, S Hooson, D Abraham, G Carta, Protein and virus-like particle adsorption on perfusion chromatography media, J Chromatogr A 1297 (2013) 96–105 [26] M.J Berry, C.E Samuel, Detection of subnanogram amounts of RNA in polyacrylamide geis in the presence and absence of protein by staining with silver, Anal Biochem 124 (1) (1982) 180–184 [27] M Strauss, N Jayawardena, E Sun, R.A Easingwood, L.N Burga, M Bostina, Cryo-electron microscopy structure of Seneca Valley virus procapsid, J Virol 92 (6) (2018) e01927 -17 [28] H.F Liu, B McCooey, T Duarte, D.E Myers, T Hudson, A Amanullah, R van Reis, B.D Kelley, Exploration of overloaded cation exchange chromatography for monoclonal antibody purification, J Chromatogr A 1218 (39) (2011) 6943–6952 14 ... stages of column loading and their abundance increased with increasing loading Hence, their further reduction would require fine-tuning of both pH and [NaCl] in the binding conditions instead of modulating... operation of the Poros 50 HS step in bind and elute mode rather than flowthrough mode for CVA21 puri? ?cation The observation that the bind and elute CEX step contributed to a further reduction of process... weaker binding one, eluting in the salt gradient, and a stronger binding one, eluting in the strip The composition of the two populations was determined via SDSPAGE analysis of pools E3 and S