1

Towards purification of adenoviral vectors based on membrane technology 1

C. Peixoto1, T.B. Ferreira1, M.F.Q. Sousa1, M.J.T. Carrondo1,2, P. M. Alves1,* 2

3

4

5

6

7

8

9

10

11

Running title: Membrane technology for adenovirus purification 12

13

14

15

1 – ITQB-UNL/IBET, Apartado 12, P-2780-901 Oeiras, Portugal; 16

2 – Laboratório de Engenharia Bioquímica, FCT/UNL, P-2825 Monte da Caparica, 17

Portugal. 18

19

20

* Corresponding author mailing address: 21

ITQB-UNL/IBET, Apartado 12, P-2780-901 Oeiras, Portugal; 22

phone: ++ 351 21 446 94 21 23

fax: ++ 351 21 442 11 61 24

e-mail: marques@itqb.unl.pt 25

26

2

Abstract 1

Given increasing applications of recombinant adenoviruses for gene therapy and 2

vaccination there is a need for highly robust and fast purification platforms for their 3

large scale manufacture. 4

Traditional chromatographic methods using resins as matrices have several 5

limitations such as high-pressure drops, slow processing rates due to pore diffusion 6

and channelling of the feed through the bed. In contrast, membrane adsorbers offer the 7

advantage of fast, gentle and effective isolation. Furthermore, membranes are easy to 8

use, no column packing is needed and, when used as disposables, no cleaning 9

validation is necessary, representing a substantial advantage to meet cGMP 10

requirements. In this work a strategy for purification of adenovirus vectors from cell 11

culture bulks fully based on membrane devices is presented. 12

Ultrafiltration membranes with molecular weight cutoffs of 300, 500 and 750 13

kDa were tested for concentration of cell culture supernatant after an initial 14

clarification step. The results show that the use of ultrafiltration/diafiltration 15

membranes not only concentrates the virus but also leads to a removal of 90% of host 16

cell DNA and proteins in the retentate. 17

Two membrane adsorbers (Sartobind Q and Sartobind Anion Direct) were 18

evaluated for adenovirus vectors capture and purification. In order to define the best 19

operating conditions, the effect of pH, conductivity and recirculation of load bulk on 20

the recovery yield of infectious adenoviruses were evaluated. Sartobind Anion Direct 21

allows for higher recovery yields (up to 62 %) of infectious adenoviruses than 22

Sartobind Q; identical ratios between total and infectious adenoviruses (TP/IP) were 23

achieved for both membrane adsorbers. 24

The overall recovery yield of the process is approximately 52%, this work 25

credit membrane technology as an alternative for the concentration and purification of 26

3

adenoviruses and as a promising solution for downstream processing of other viral 1

vectors. 2

3

Key words: adenovirus vectors, downstream processing, membrane adsorber, anion-4

exchange, purification 5

6

7

8

9

10

11

12

13

14

15

16

17

18

19

20

21

22

23

24

25

26

4

Introduction 1

The use of recombinant adenoviruses for vaccination and gene therapy 2

requires fast and highly efficient purification protocols that allow for the maintenance 3

of viral infectivity, high recoveries of infectious particles, removal of contaminating 4

DNA and host cell proteins while also permitting to concentrate the viral samples for 5

final delivery. Concentration of adenoviruses is critical to obtain high titer vector 6

stocks and to decrease the handling volume thus simplifying the ensuring purification 7

steps. Process scalability is also a prerequisite for large scale production of 8

adenoviruses for vaccines or gene therapy (Bauer, Pilaro et al. 2002; Tatsis and Ertl 9

2004). 10

A number of chromatography-based methodologies have been reported for 11

purification of adenoviruses (Green, Huang et al. 2002; Arcand, Bernier et al. 2003; 12

Kamen and Henry 2004; Lusky 2005), anion exchange chromatography being the 13

driving force for most of the processes described in the literature (Altaras, Aunins et 14

al. 2005). However, most chromatographic media exclude adenoviruses due to their 15

pore size distribution, leading to very low binding capacities. Indeed, as media 16

transport of solutes into the porous lumen of beads occurs by diffusion, for large 17

biomolecules, even for pores larger than 100 nm, mass transfer is still limiting 18

(Thoemmes and Kula 1995; Charcosset 2006). Furthermore, resins require column 19

packing, long regeneration times and the packed resin is susceptible to fouling (Ghosh 20

2002). 21

Latest improvements in membrane materials and chemistries have generated 22

renewed interest in applications of membrane chromatography for bioprocessing 23

(Charcosset 2006; van Reis and Zydney 2007). Membrane adsorbers work on the 24

same principles as resin-based chromatography systems, i.e. by binding of the ligand 25

5

to the membrane. The major difference between membrane and resin supports is the 1

porosity; the larger pores of anion exchange membranes strongly reduce diffusion 2

limitations, with concomitant improvement of flow rates and chromatography 3

kinetics. Furthermore, membrane chromatography can be used as a disposable device, 4

without the need for washing, cleaning or regeneration, significantly shortening the 5

processing time, decreasing buffer volumes and lessening validation issues when 6

working under a GMP environment. 7

Membrane adorbers have been used as alternatives to column chromatography 8

for the final polishing step (removal of endotoxins and viruses) in antibody 9

production (Zhou and Tressel 2006) as well as for plasmid DNA purification (Syren, 10

Rozkov et al. 2007). Several examples were described for virus capture and/or 11

purification using membrane adsorbers (Karger, Bettin et al. 1998; Specht, Han et al. 12

2004; Kalbfuss, Wolff et al. 2007) and tangential flow filtration of adenovirus using 13

hollow fibers has previously been reported by our group (Peixoto, Ferreira et al. 14

2006); the relevance of membrane pore size under process conditions has also been 15

established (Subramanian, Altaras et al. 2005). Moreover, membrane chromatography 16

was applied for adenoviruses purification (Duffy, O'Doherty A et al. 2005; Lajmi, 17

Kutner et al. 2007). Nevertheless, a complete purification process addressing only 18

membrane technology has not been reported yet. 19

This work described the development of a fast, robust and scaleable 20

purification process solely based on membrane technology, fulfilling the requirements 21

for disposable technologies in the biopharmaceutical industry. The process hereby 22

described and developed is divided into three parts: (i) Clarification by dual 23

microfiltration and concentration by cross-flow ultrafiltration, followed by (ii) 24

purification using anion exchange membrane adsorbers and (iii) concentration and 25

6

buffer exchange for final delivery formulation. Similar downstream processes might 1

be applicable to less prevalent adenoviruses serotypes which are emerging (Konz, 2

Livingood et al. 2005; Stone and Lieber 2006), with slight modifications. 3

4

Materials and Methods 5

Cell line and medium 6

293 cells (purchased from ATCC (ATCC-CRL-1573)), were adapted to 7

suspension and grown in shake flasks (Corning, USA) in CD293 supplemented with 4 8

mM of glutamine (all from Invitrogen, UK) in a humidified atmosphere of 8% CO2 in 9

air at 37ºC. Cells were routinely propagated twice a week using an inoculum of 10

0.5×106 cell/ml. Cell concentration and viability were determined by counting cells on 11

a Fuchs-Rosenthal haemocytometer (Brand, Germany) using the trypan blue 12

(Invitrogen, UK) dye exclusion method. 13

14

Virus production 15

A replication-defective adenovirus type 5 (Ad5) expressing GFP protein was 16

used. For all experiments virus production was performed in a 5 Liter bioreactor 17

(Sartorius, B.Braun, Germany). The dissolved oxygen and pH were controlled at 80% 18

of air saturation and 7.2, respectively. Infections were done using a MOI of 10; for the 19

downstream experiments the adenoviruses were harvested at 72 hours post infection 20

(hpi). The amount of infectious particles in the bulk (before purification) was in the 21

range of 109 to 1010 IP/ml. 22

23

Quantitation of infectious particles 24

Adenovirus titration was done according (Cote, Bourget et al. 1997). Briefly, 25

7

anchorage-dependent 293 cells were culture in 24 well plates at 0.25×106 cell/ml. One 1

day later, cells were infected with 1 ml of serial dilutions of the adenoviruses in 2

infection medium (DMEM + 10% (v/v) FBS). After 17 to 20 h of incubation the 3

supernatant was collected and the cells were trypsinised and harvested. Cell 4

suspensions were immediately analyzed in a flow cytometer CyFlow® Space (Partec, 5

Germany) equipped with a blue solid state laser set at 488 nm and data analysis was 6

performed using FlowMax® software (Partec, Germany). 7

8

Measurement of total adenovirus vectors by quantitative real time PCR 9

Total adenoviruses were quantified by real time PCR. Viral DNA was 10

extracted and purified with the “High Pure Viral Nucleic Acid Kit” (Roche 11

Diagnostics, Germany). Purified adenoviral DNA previously purified and quantified 12

(Peixoto, Ferreira et al. 2006) was used as external standard to quantify adenovirus 13

type 5. 14

The size of the amplicon is 600 bp. DNA amplifications were done using the Light 15

Cycler system with “Fast Start Master SYBR Green I kit” (Roche Diagnostics, 16

Germany). For a 20 µl PCR reaction, 2 µl DNA template and 18 µl of mastermix were 17

added to each capillary. The mastermix was prepared to obtain a final concentration 18

of 2 mM MgCl2 and 0.5 µM of each primer in each capillary. The oligonucleotide 19

primers used were: 20

The presence of adenovirus serotype 5 (Ad5) was detected by the forward (5' 21

GAAAAGGCTGTCCGTGTCC 3´) and reverse (5' CTAAACATCGCGGCCAAG 3´) 22

primers, designed to amplify a fragment of the E1A gene of 600 bp. DNA was 23

denatured for 10 min at 95ºC. The standard amplification program included 40 three-24

step cycles: (i) denaturation-heating at 95ºC for 1 second, (ii) annealing-cooling at 25

8

60ºC during 8 seconds and (iii) elongation-heating at 72ºC for 24 seconds. 1

Fluorescence was acquired at the end of the elongation step of each cycle. Data was 2

analyzed using the Light Cycler software and the “fit points method” described 3

elsewhere (Rasmussen, Morrison et al. 1998). 4

5

Downstream Process 6

Clarification and concentration 7

To promote cell lysis, 0.1% Triton X-100 (Merck, Germany) was added to the 8

bulk at the end of bioreaction (72 h); the bioreactor agitation rate was increased to 9

1000 rpm for 1 minute; afterwards the bulk was incubated for 30 min at 37 ºC with 10

Benzonase (15U/ml) (Merck, Darmstadt). Clarification was carried out by 11

microfiltration (MF) with a flow rate of 90 ml/min using Sartobran 0.45µm filter 12

(Sartorius AG, Germany) with 100 cm2 of filtration area or with a flow rate of 70 13

ml/min using Sartopure 2 polyethersulfone membrane capsules (Sartorius AG, 14

Germany), with a pre-filter of 0.8 µm followed by a 0.45 µm filter and with 50 cm2 of 15

filtration area. 16

The clarified bulks were concentrated by tangential flow filtration using 17

polysulfone hollow-fiber cartridges with of 750, 500 and 300 kDa cut-offs and 420 18

cm2 of total filtration area each one as well as 1 mm of lumen diameter fibers (GE 19

Healthcare, USA). The cartridges were coupled to QuixStand apparatus (GE 20

Healthcare, USA) and the recirculation flow rate used with the hollow fiber cartridges 21

was constant and equal to 600 ml/min, with continuos recirculation of the concentrate 22

to the feed with no backpressure on the permeate until rentante volume reached 10% 23

of the initial feed volume. 24

25

Anion exchange membrane chromatography 26

9

Preliminary anion exchange experiments were performed with Sartobind Q 1

MA75 and Sartobind Anion Direct (2.5 ml) provided by Sartorius AG. The 2

membranes were coupled to an Akta Explorer chromatography system (GE 3

Healthacare, USA). UV, conductivity and pH were monitored on line. 4

Both membrane adsorbers were equilibrated with 10 mM Tris-HCl pH 8.0 5

before loading with 50 ml of concentrated, clarified bulk (previously concentrated and 6

clarified). The pH of the bioreactor bulk was adjusted by addition of NaOH and/or by 7

diafiltration with Vivaspin 20 with a 100 kDa cut-off (Sartorius AG, Germany) using 8

appropriate buffer capacities (10 mM Tris-HCl pH 8.0). The loading of the 9

membranes was performed using a flow rate of 3 ml/min or by recirculation for 60 10

min with a flow rate of 20 ml/min; capture of adenoviruses was monitored by 11

recovering samples every 10 min up to one hour. Adsorbed viruses were displaced by 12

a linear salt gradient increasing to 1 M NaCl or by a salt step (0.2 M, 0.65 M and 1 M 13

NaCl). Samples of the feed, flowthrough and eluate were collected and analyzed for 14

total and infectious adenoviruses particles. 15

To evaluate scaleability a membrane capsule of Sartobind Anion Direct with 16

900 cm2 of area was used. The membrane was coupled to Quixstand system and fed 17

with a recirculation flow rate of 200 ml/min. Membrane adsorbers were cleaned with 18

1 M NaOH between runs and rinsed with 20% ethanol before storage at 4ºC. 19

After purification with membrane adsorbers, the adenoviruses were further 20

concentrated using polysulfone hollow-fiber cartridge with 300 kDa cut-off and 420 21

cm2 of filtration area as well as 1 mm of lumen diameter fibers (GE Healthcare, 22

USA) ultrafiltration was performed using a recirculation flow rate of 600 ml/min with 23

no backpressure on permeate side until concentrated volume reach 10% of the initial 24

feed volume. A sterile microfiltration with Filtropur S 0.2 µm (Sarstedt, Germany) 25

was performed before storage at 4ºC. 26

10

1

Analysis by size exclusion chromatography 2

The purified adenoviruses were analyzed by size exclusion chromatography using a 3

prepacked Hiprep 16/60 Sephacryl S-300 column (1.6 cm of diameter and 60 cm of 4

length) coupled to an AKTA explorer (both GE Healthcare, USA). The elution buffer 5

was 10 mM Tris-HCl pH 8.0 used with a flow rate of 54 cm/h. 6

7

Protein and DNA analysis 8

Protein analysis by sodium dodecyl sulfate/polyacrylamide gel electrophoresis 9

(SDS-PAGE) was performed with 4-12% NuPage (Invitrogen, UK) gradient precast 10

gels using MES running buffer (Invitrogen). Protein bands were visualized by Simply 11

Blue Safe Stain (Invitrogen, UK). The total protein in the samples for loading the 12

SDS-PAGE were quantified using a BCA kit 23225 (Pierce, USA) on 96 well plates. 13

Western blot analysis was carried out using an Invitrogen PAGE and semi dry transfer 14

unit (Amersham Biosciences, USA) and the proteins of interest detected with a 15

polyclonal rabbit anti-adenovirus type 5 (Abcam, UK) in combination with alkaline 16

phosphatase conjugated anti-rabbit immunoglobulin G (IgG) (Sigma, MO) using 17

BCIP/NBT substrate (Pierce, USA). 18

Quantification of residual DNA content was undertaken by PicoGreen dsDNA Assay 19

Kit (Molecular Probes, USA). 20

21

Endotoxins quantitation 22

The endotoxin level in the end-products was determined by LAL test (Limulus 23

Amebocyte Lysate QCL-1000 kit, Cambrex). 24

25

26

11

Results and Discussion 1

Clarification and Concentration 2

Cell lysis was performed by addition of 0.1% Triton X-100 followed by 3

clarification using membrane capsules with a pre-filter of 0.8 µm and a filter with 4

0.45 µm pores resulting in a recovery yield of infectious particles of 98±5%. The 5

introduction of a pre-filter prevents clogging of the membranes due to entrapment of 6

the cell debris and adenoviruses in the membrane pores and results in higher 7

recoveries of adenoviruses in the filtrate. Taking into account this result, the harvest 8

and clarification operating conditions were used for further experiments. 9

Concentration of adenoviruses was performed using tangential flow filtration; 10

the influence of membrane pore size in recovery yields of adenoviruses is shown in 11

Table 1. An improvement of 54 % in recovery yield was obtained when pore size was 12

reduced from 750 kDa to 300 kDa. Pore sizes larger than 300 kDa allowed for the 13

passage of the adenoviruses to the permeate, decreasing the number of infectious 14

particles in the retentate. Nevertheless, a three fold increase in process time was 15

achieved. 16

17

Purification using membrane adsorbers 18

To remove remaining contaminating proteins and DNA from concentrated 19

adenoviral stocks another purification step was performed with membrane adsorbers, 20

namely Sartobind Q and Sartobind Anion Direct Q. The influence of pH on the 21

binding and elution of adenoviruses from membrane adsorbers was investigated 22

(Table 2). Four samples of 50 ml, collected from the same bioreactor bulk, were 23

adjusted to different pHs, clarified, concentrated and adjusted for a similar 24

conductivity (2.7 mS) (as described in Materials and Methods) before loading onto 25

12

Sartobind Q 75. The highest yield was obtained at pH 8.0 (aprox. 43 % of virus 1

recovered); thus a constant pH of 8.0 was used for all subsequent steps. 2

To determine the elution profiles for the membrane adsorbers, linear gradient 3

elutions were performed with increasing salt concentrations (NaCl). These profiles 4

allow for the determination of the separation capacity of the media and the 5

conductivity at which the adenoviruses were displaced from the membrane. 6

Adenovirus eluted in the range of 35-60 mS/cm conductivity with a concentration of 7

7.2 ×108 TP/ml, whereas the majority of the bound contaminants eluted at 15-20 8

mS/cm (Fig. 1A). Based on this chromatographic profile, elution of the adenoviruses 9

using NaCl steps was attempted in order to achieve concentration factors compatible 10

with final dose formulation (Raper, Chirmule et al. 2003) and to reduce process time. 11

Three steps were defined: (i) 250 mM NaCl to elute protein contaminants; (ii) 650 12

mM to elute the adenoviruses; (iii) a third step with 1 M NaCl to wash the membrane 13

and displace the DNA (Fig. 1B); the second step was not performed at higher 14

conductivity to avoid contamination of the adenoviruses fraction with remaining 15

DNA. 16

The total amount of adenoviruses detected in flowtrough for both elution 17

modes is negligible (aprox.105 TP/ml) in respect to the final recovery yields of 18

adenoviruses. Nevertheless, the concentration factor obtained with step elution is at 19

least 6 fold higher than the concentration obtained for NaCL gradient elution (total 20

adenoviruses in the middle peak is 4.5×109 TP/ml). Futhermore, a substantial 21

reduction in process time was also achieved for step gradient elution; however, the 22

yield of infectious particles in the pooled fraction was low (30-40%). Finally, it was 23

observed that variations in feed conductivity, up to 15-20 mS/cm, do not significantly 24

affect the recovery of infectious adenoviruses after elution, as determined after 25

loading adenoviruses samples from different batches (data not shown). 26

13

In an attempt to improve the process yield, the effect of feed recirculation on 1

recovery yield was evaluated. The previously concentrated sample was loaded in 2

recirculation mode for 60 min, the capture profile for Sartobind Q and Sartobind 3

Anion Direct Q (Figure 2 A and 2 B) indicate that Sartobind Anion Direct Q is more 4

efficient to capture adenoviruses than Sartobind Q. After 30 min in recirculation mode 5

only negligible amounts of infectious particles were found in flow through of the 6

Sartobind Anion Direct Q (typically less than 5% of loaded adenoviruses titer), 7

whereas comparable results were obtained for Sartobind Q only 20-30 minutes later. 8

The Sartobind Anion Direct was shown to be more suitable for the purification 9

of adenoviruses. The results also confirm that the capacity of the membranes was not 10

exceeded as the first results obtained without feed recirculation (30 - 40% of 11

recovery) could indicated. The determined capacity is in the order of 1 to 1.5×1013 12

viral particle/ml of membrane volume (3 to 4×1011 viral particle/ cm2). The recovery 13

yield obtained after recirculation of the feed for both Sartobinds is approximately 14

20±8 % higher than the recovery achieved without recirculation; sample loading in 15

recirculation mode improved the intermediate purification step due to a longer 16

residence time of the adenoviruses. 17

Sartobind Anion Direct was used exclusively for subsequent scale-up studies 18

due to its higher capture capacity per unit of time leading to a significant reduction in 19

process time, which may impact the robustness of the process. A 10-fold scale-up of 20

this downstream process based on membrane technology was performed in order to 21

test the robustness, scaleability of the process and further substantiate product quality. 22

The optimized operating conditions were applied in a simple process where 23

the feed was driven by a peristaltic pump and the residence time was kept constant. 24

The results and main parameters (Table 3) are in agreement with the small-scale 25

results. The total virus to infectious unit (TP/IP) ratio, a well-accepted standard 26

14

assessement and requirement for quality control of any adenoviral vector is an 1

indicator of product potency (McIntyre, 2001; Adenovirus Reference Material 2

Working group). The ratios obtained after the Sartobind Anion Direct stay below 10, 3

substantially lower than the maximum FDA required ratio of 30:1. 4

The purified adenoviruses sample was also analysed by gel electrophoresis 5

(Fig. 3); the band pattern is comparable to adenoviruses purified by other methods as 6

described in the literature (Arcand, Bernier et al. 2003). DNA level and endotoxin 7

content are in the same order of magnitude as those obtained for the small scale 8

experiments, i.e. endotoxins are below 10 EU/ml whereas DNA contaminantion was 9

reduced 90% after the ultrafiltration step; and the amount of DNA quantified after the 10

membrane adsorber purification is 13 ng per 1011 TP, meaning less than 3% of the 11

initial amount of DNA remains. However, dependending of the potency determined, 12

since the FDA specifications for adenovirus vectors (DNA content below 10 ng/dose) 13

is very restricted (Bauer, Pilaro et al. 2002), additional operation specifically aimed at 14

reducing DNA might be incorporated as increasing the amount of nuclease added in 15

the beginning of downstream process or DNA precipitation (Altaras, Aunins et al. 16

2005; Goerke, To et al. 2005). 17

After the high purity achieved with the three membrane purification steps, a 18

diafiltration/concentration step was introduced next to the anion exchange membrane, 19

to obtain the adenoviruses in the desired concentration for administration and storage. 20

The recovery yield obtained is 90% (Table 4), slightly lower than that obtained in the 21

ultrafiltration performed before purification with membrane adsorber; this can be due 22

to shearing stress effects and/or non-specific adsorption onto the membrane as the 23

viruses at this stage are much purer than at the beginning of the purification process. 24

The eventual need for a final polishing step was evaluated using a size 25

exclusion chromatography column. The chromatographic profile obtained was 26

15

compared with the size exclusion chromatography analysis performed immediately 1

before the intermediate purification using Sartobind Anion Direct Q (Fig.4 A), where 2

several peaks corresponding to contaminant proteins were eluted at higher volumes 3

than adenoviruses. The chromatogram profile of the final product shows only one 4

peak corresponding to adenoviruses, with a UV detection ratio of 260 nm over 280 5

nm of 1.2. (Fig.4 B). 6

These results indicate that size exclusion is not necessary for final polishing of 7

the adenoviruses. Moreover, undesirable dilution of the final product would take place 8

if size exclusion chromatography had been used, a dilution factor of 10-15 being 9

observed. A recovery of 99±5 % was obtained after final microfiltration, this step 10

assuring sterility of the final product. 11

Integrating results obtained here with the dvances for non-Ad5-based vectors 12

(Stone and Lieber 2006) together with the possibility to correlate retention behaviour 13

with electrostatic properties of the hexon protein (Konz, Livingood et al. 2005) will 14

permit the design of a platform for purification of adenovirus vectors. 15

16

Concluding remarks 17

The scaleable purification process described herein comprises four steps based 18

on membrane technology fitting the trends for utilization of disposables. An 19

intermediate recovery yield up to 62% was obtained using Sartobind Anion Direct 20

after optimization of operating parameters, the contaminant DNA showing over 97 % 21

reduction whereas the endotoxin contents stay below 10 EU/ml. This purification 22

process represents a substantial improvement and is an appealing alternative to 23

chromatographic resins based methods, as it reduces downstream process time 24

without compromising the yields of adenoviruses for clinical trials. 25

26

16

Acknowledgements 1

The authors thank Rosário Clemente for technical support. Sartorius is also 2

acknowledged, in particular Dr. Noushin Delmdahl, for technical advice and for 3

providing the membranes. The authors acknowledge and appreciate the financial 4

support received from the European Commission (Project ICF599A4PR01) and from 5

Fundação para a Ciência e Tecnologia – Portugal (Project POCTI/BIO/46515/2002 6

and SFRH/BD/10614/2002). 7

8

17

Figure Captions 1

2 Figure 1 – Linear gradient (A) and step (B) elution profiles obtained after loading the 3

membrane adsorber (Sartobind Q75) with 50 ml of previously clarified and 4

concentrated adenoviral supernatants in 10 mM Tris-HCL pH 8.0; loading and elution 5

were performed at flow rates of 3 ml/min. The arrow indicates the peak corresponding 6

to total adenoviruses particles. 7

8 Figure 2 – Adenoviruses capture profile on Sartobind Q (A) and Sartobind Anion 9

Direct Q (B) using recirculation mode for 60 minutes; the samples were loaded at 20 10

ml/min. Co is the initial amount of infectious adenoviruses and C is the amount of 11

infectious adenoviruses at sampling time. 12

13 Figure 3 – - SDS-PAGE (A) and Western Blot (B) analysis samples derived from 14

adenoviruses purification. SDS-PAGE (A): analysis of samples collected during 15

purification using membrane technology. The gel was loaded with 25 µg of total 16

protein per well excepted lane 7 that was loaded with 100 µg of total protein. 1 - MW 17

markers (kDa); 2 - adenoviral supernatant after clarification and concentration by 18

ultrafiltration; 3 - ultrafiltration permeate; 4 - flow trough of membrane adsorber 19

(Sartobind Anion Direct); 5 - membrane wash before elution; 6 - MW markers (kDa); 20

7 - peak eluted at 650 mM .Western Blot (B): 1 – prestained molecular weight 21

markers; 2 - CsCl purified adenovirus 3 - peak eluted at 650 mM. 22

23

Figure 4 – Quality assessment of adenoviruses using size exclusion chromatography. 24

Chromatographic profile obtained before intermediate purification step (A) and after 25

purification using a Sartobind Anion Direct (B).26

18

Table 1: Effect of membrane cut-offs on the recovery yield of infectious adenoviruses 1

during tangential flow filtration 2

3

4 5 6 a) Recovery was calculated as the ratio between infectious adenoviruses titers of final concentrate and 7 infectious adenoviruses in the initial supernatant. 8 9 10 11 12 13 Table 2: Effect of loading pH on the recovery yield of infectious adenoviruses 14 purified with membrane adsorbers (Sartobind Q). 15 16

pH of Load materialª 6.5 7.4 8.0 8.5

Recovery yield of IP (%) 15 14 43 37

17 a) Four samples of 50 ml collected from the same bioreactor bulk were clarified, concentrated and 18 adjusted for conductivity. The infectious adenoviruses titer was measured by flow cytometry as 19 described in Materials and Methods. 20 21 22 23 24 Table 3: Scale-up of the purification using Sartobind Anion Direct Q membranes. 25 26

Membrane adsorber

Membrane area (cm2)

Flow (L/min)

Residence time (min )b

Total adenoviruses

load (TP) TP/IP c

Yield of infectious adenoviruses recovery

(%) Sartobind anion direct 2.5 ml a

90 0.02 0.125 2.9×1011 6±4 60±10

Sartobind anion direct 25 ml a 900 0.20 0.125 4.0×1014 9±3 62±2

27 a) The Sartobind anion direct with 2.5 ml of membrane volume was coupled to an AkTA Explorer 28 whereas the Sartobind anion direct of 25 ml was connected with appropriate tubing to the QuixStand 29 System; the feed was driven with a peristaltic pump. 30 31 b) The residence time was kept constant for both membranes during the feed, wash and elution. 32 33 c) TP/IP – Ratio between total and infectious adenoviruses is an indicator of product potency and 34 should be 30 or less, as required by FDA 35

NMW (kDa)

Process Time (min)

Infectious adenoviruses

(IP) in retentate

Infectious adenoviruses (IP)

in permeate

Recoveryª yield of infectious adenoviruses

(%)

750 20 3.0×109 2.9×109 42 500 45 6.0×1010 2.2×109 71 300 67 1.5×1010 1.2×106 96

19

1 Table 4: Evaluation of downstream process steps for adenoviruses purification using 2 membrane technology. 3

4 5

Process Step TP/IP b ng DNA / 1011TP

Endotoxins

(EU/ml)

Recovery yield of IP a (%)

Clarification 167 - - 98±5

Ultrafiltration 50 - - 96±2

AEX membrane 12 - - 62±2

Ultrafiltration - - - 90±3

Sterile filtration 12 - - 99±5

Final Product 12 13 < 10 52±5

6 a) IP – infectious adenoviruses. 7 b) TP/IP – Ratio between total and infectious adenoviruses is an indicator of product potency and 8

should be 30 or less, as required by FDA. 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40

20

Figure 1 - 1 2 3 4 A 5 6

0

20

40

60

80

100

120

0 20 40 60 80 100 120 140 160

Volume (ml)

Ab

s 28

0 n

m (

mA

U)

0

20

40

60

80

100

Co

nd

uct

ivit

y (m

S/c

m)

Abs 280 nm

Conductivity

Ad

7 8 B 9

10 11

0

500

1000

1500

2000

2500

3000

0 20 40 60 80 100 120 140 160 180 200

Volume (ml)

Ab

s 28

0 n

m (

mA

U)

0

20

40

60

80

100

Co

nd

uct

ivit

y (m

S/c

m)

Abs 280 nm

ConductivityAd

12 13

21

Figure 2 - 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26

27 28

29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51

A - A - Sartobind Q

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 10 20 30 40 50 60

time (min)

C/C

o

B - Sartobind Direct Q

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 10 20 30 40 50 60

time (min)

C/C

o

A - A - Sartobind Q

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 10 20 30 40 50 60

time (min)

C/C

o

B - Sartobind Direct Q

0,0

0,2

0,4

0,6

0,8

1,0

1,2

0 10 20 30 40 50 60

time (min)

C/C

o

22

1 Figure 3 A – 2 3 4

5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 Figure 3 B – 20 21

22

1 2 3 4 5 6 7kDa

9766

45

31

20

14

1 2 3 4 5 6 7kDa

9766

45

31

20

14

1 2 3

45

Hexon

III

core

23

1 2 Figure 4 - 3 4 A 5

0

100

200

300

400

500

0 20 40 60 80 100 120 140

volume (ml)

Ab

s 28

0 n

m (

mA

U)

6 7 8

9 B 10 11

0

100

200

300

400

500

600

0 20 40 60 80 100 120 140

volume (ml)

Ab

s 28

0 n

m (

mA

U)

12 13

24

Bibliography 1

Altaras, N. E., J. G. Aunins, et al. (2005). "Production and formulation of adenovirus 2

vectors." Adv Biochem Eng Biotechnol 99: 193-260. 3

Arcand, N., A. Bernier, et al. (2003). "Adenovirus type 5 (Ad5)chromatographic 4

purification process at 20L scale." Bioprocess J 2: 72-75. 5

Bauer, S. R., A. M. Pilaro, et al. (2002). Testing of Adenoviral Vector Gene Transfer 6

Products: FDA Expectations. Adenoviral Vectors for Gene Theraphy. D. T. 7

Curiel and J. T. Douglas, Academic Press: 615-654. 8

Charcosset, C. (2006). "Membrane processes in biotechnology: An overview." 9

Biotechnol Adv 24(5): 482-492. 10

Cote, J., L. Bourget, et al. (1997). "Study of adenovirus production in serum-free 11

293SF suspension culture by GFP-expression monitoring." Biotechnol Prog 12

13(6): 709-14. 13

Duffy, A. M., M. O'Doherty A, et al. (2005). "Purification of adenovirus and adeno-14

associated virus: comparison of novel membrane-based technology to 15

conventional techniques." Gene Ther 12 Suppl 1: S62-72. 16

Ghosh, R. (2002). "Protein separation using membrane chromatography: opportunities 17

and challenges." J Chromatogr A 952(1-2): 13-27. 18

Goerke, A. R., B. C. To, et al. (2005). "Development of a novel adenovirus 19

purification process utilizing selective precipitation of cellular DNA." 20

Biotechnol Bioeng 91(1): 12-21. 21

Green, A. P., J. J. Huang, et al. (2002). "A new scalable method for the purification of 22

recombinant adenovirus vectors." Hum Gene Ther 13(16): 1921-34. 23

Kalbfuss, B., M. Wolff, et al. (2007). "Direct capture of influenza A virus from cell 24

culture supernatant with Sartobind anion-exchange membrane adsorbers." J 25

Membrane Sci 299(1-2): 251-260. 26

Kamen, A. and O. Henry (2004). "Development and optimization of an adenovirus 27

production process." J Gene Med 6 Suppl 1: S184-92. 28

Karger, A., B. Bettin, et al. (1998). "Simple and rapid purification of 29

alphaherpesviruses by chromatography on a cation exchange membrane." J 30

Virol Methods 70(2): 219-224. 31

Konz, J. O., R. C. Livingood, et al. (2005). "Serotype specificity of adenovirus 32

purification using anion-exchange chromatography." Hum Gene Ther 16(11): 33

1346-53. 34

25

Lajmi, A. R., R. Kutner, et al. (2007). A membrane chromatography application: a 1

rapid, high capacity gene therapy vector purification tool. Process scale 2

bioseparations for the biopharmaceutical industry. NW, Taylor & Francis. 31: 3

541. 4

Lusky, M. (2005). "Good manufacturing practice production of adenoviral vectors for 5

clinical trials." Hum Gene Ther 16(3): 281-91. 6

Peixoto, C., T. B. Ferreira, et al. (2006). "Purification of adenoviral vectors using 7

expanded bed chromatography." J Virol Methods 132(1-2): 121-126. 8

Raper, S. E., N. Chirmule, et al. (2003). "Fatal systemic inflammatory response 9

syndrome in a ornithine transcarbamylase deficient patient following 10

adenoviral gene transfer." Mol Genet Metab 80(1-2): 148-58. 11

Rasmussen, R., T. Morrison, et al. (1998). "Quantitative PCR by continuos 12

fluorescence monitoring of a double strand DNA specific binding dye." 13

Biochemica No.2: 8-11. 14

Specht, R., B. Han, et al. (2004). "Densonucleosis virus purification by ion exchange 15

membranes." Biotechnol Bioeng 88(4): 465-73. 16

Stone, D. and A. Lieber (2006). "New serotypes of adenoviral vectors." Curr Opin 17

Mol Ther 8(5): 423-31. 18

Subramanian, S., G. M. Altaras, et al. (2005). "Pilot-Scale Adenovirus Seed 19

Production through Concurrent Virus Release and Concentration by Hollow 20

Fiber Filtration." Biotechnol Prog 21(3): 851-9. 21

Syren, P.-O., A. Rozkov, et al. (2007). "Milligram-scale parallel purification of 22

plasmid DNA using anion exchange membrane capsules and a multi-channel 23

peristaltic pump." J Chromatogr B in Press. 24

Tatsis, N. and H. C. J. Ertl (2004). "Adenoviruses as Vaccine Vectors." Mol Ther 25

10(4): 616-629. 26

Thoemmes, J. and M.-R. Kula (1995). "Membrane chromatography - an integrative 27

concept in the downstream processing of proteins." Biotechnol. Prog. 11(4): 28

357-367. 29

van Reis, R. and A. Zydney (2007). "Bioprocess membrane technology." J Membrane 30

Sci 297(1-2): 16-50. 31

Zhou, J. X. and T. Tressel (2006). "Basic Concepts in Q Membrane Chromatography 32

for Large-Scale Antibody Production." Biotechnol. Prog. 22(2): 341-349. 33

34