Adenovirus as a Biopharmaceutical

Tiago Bruno Ferreira Adenovirus as a Biopharmaceutical: Optimisation of vector production Adenovirus as a Biopharmaceutical Dissertation presented to obtain a Ph.D degree in Chemical Engineering at the Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa Tiago Bruno Ferreira Optimisation of vector production Oeiras, March 2007 Vaccination has been one of the most important interventions designed to prevent disease to be employed on a worldwide basis. Adenoviral vector has been extensively studied as a vaccine platform because of its ability to induce potent cellular and humoral immunity. Consequently, the market requirements for adenoviral vectors are increasing, creating a need for new methodologies for largescale production of concentrated vectors with warranted purity and efficacy. This work contributes for the improvement of the state of the art of adenoviral vectors production process. Adenovirus as a Biopharmaceutical Optimisation of vector production By Tiago Bruno Pereira Soares Ferreira Dissertation presented to obtain a Ph.D degree in Chemical Engineering at the Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa Oeiras, March 2007 Adenovirus as a Biopharmaceutical Optimisation of vector production Tiago Bruno Pereira Soares Ferreira Dissertation presented to obtain a Ph.D degree in Chemical Engineering at the Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa Supervisors: Paula M. Alves and Manuel J. T. Carrondo Instituto de Tecnologia Química e Biológica Universidade Nova de Lisboa Oeiras, March 2007 ii Adenovirus as a Biopharmaceutical: Optimisation of vector production Second edition: May 2007 Copyright number: Cover images: Front: Adenovirus particle. Source: image was adapted from the web site http://nobelprize.org/nobel_prizes/medicine/laureates/1993/illpres/gene s-in-pieces.html. Back: Electron microscope picture of adenovirus particles (up image) and microscope picture of 293 cells infected with adenovirus vectors (back image). ITQB/IBET Animal Cell Technology Laboratory Instituto de Tecnologia Química e Biológica/ Instituto de Biologia Experimental e Tecnológica Apartado 12, 2781-901 Oeiras, Portugal Fax: +351 21 442 11 61; Phone: +351 21 446 94 24 http://www.itqb.unl.pt http//www.ibet.pt iii Supervisors: Dr. Paula Marques Alves, Principal Researcher and Group Leader of the Animal Cell Technology Laboratory at ITQB, Oeiras Portugal. Prof. Dr. Manuel J. T. Carrondo, Professor of Chemical and Biochemical Engineering at Universidade Nova de Lisboa, Monte da Caparica, Portugal, CEO (Chief Executive Officer) of IBET, Oeiras, Portugal, CSO (Chief Scientific Officer) of GenIBET Biopharmaceuticals, Oeiras, Portugal and Head of the Animal Cell Technology group at ITQB, Oeiras, Portugal. Jury: Dr. John G. Aunins, visiting Professor at ITQB, Oeiras, Portugal and Senior Director of Cell Culture Development and Vaccine BioProcess at Merck Research Labs., Philadelphia, USA. Dr. Eric J. Kremer, Group Leader and Director of Research at Institut de Génétique Moléculaire de Montpellier CNRS, Montpellier, France. Dr. Joaquim Sampaio Cabral, Professor at Instituto Superior Técnico, Lisboa, Portugal. Dr. Ana Sofia Coroadinha, Auxiliar Resercher of the associated laboratory ITQB/IGC/IBET, Oeiras, Portugal. iv Foreword The present thesis dissertation is the result of four years of research at the Laboratory of Animal Cell Technology at ITQB/IBET (Portugal) under Dr. Paula M. Alves and Prof. Dr. Manuel J. T. Carrondo Supervision. This thesis intends to explore the limiting factor on adenovirus vector production at high cell densities aiming the understand and improvement of the current inefficient process of adenovirus vector production in order to respond to the increasing demand of methodologies for the production in large scale of concentrated vectors with warranted purity and efficacy. v Acknowledgements I would like to express my gratitude to all the people who directly or indirectly supported and contributed to make possible the work presented in this thesis. A very special thanks to my supervisor Dr. Paula Marques Alves, who truly made a difference in my life. It was under her tutelage that I developed a focus and became interested in the animal biotechnology field. She provided me with direction, suggestions and support. It was through her persistence, understanding and kindness that I completed this work. I doubt that I will ever be able to convey my appreciation fully. Thank you very much. I would like to express my gratitude to my supervisor, Prof. Manuel Carrondo, whose expertise, understanding and patience added considerably to my graduate experience; for trusting and giving me the opportunity to work in the animal biotechnology field. I thank him for the guidance, support, patient, rigor, pragmatism, motivation and for “pulling the muscle to its maximum”. I thank you for this great journey that allowed me to grow as a scientist and as a person. To Dr. Pedro E. Cruz for all the support, motivation, discussions and all that he taught me. I am also extremely grateful to Dr. John Aunins and Dr. Otto-Wilhelm Merten for their input. To my “office” lab colleagues Ana Sofia Coroadinha and Isabel Marcelino for their daily conversations and all the instances in which their friendship and good mood helped me along the way. vi To Ana Lúcia Ferreira and Ricardo Perdigão, the (former) diploma students who greatly contributed for the work presented in this thesis, for their commitment, effort and teamwork. I would like to thank a few people specifically. António Roldão and Marcos Sousa for all the good moments inside and outside the lab. To Carina Ferreira Silva, always ready to help and for being part of the “Ferreira team”. To Cristina Peixoto for everything she taught me about adenovirus purification, for her support and friendship. To Change Zhang for all the support. Marlene Carmo for her patience and for all the help concerning the flow cytometric analyzes. To Rosário Clemente for her technical Support. I sincerely thank all of the current and former members of the Animal Cell Technology Lab for their help and friendship over the last four years. I really spent a great deal of time at lab. I would also like to thank my parents and brother for the support they provided me through my entire life without whose love and encouragement I would not have finished this thesis. I must give immense thanks to my patient and loving wife Alexandra and our daughter Carolina, who have been a great source of strength all through this work, especially during those times when the research came first. Their love and support was of immeasurable value to me. To the entities that funded the research: European Commission (Project RP/PPR ORALVAC ICA4-CT-2000-30027), Fundo Social Europeu (FSE) and Fundação para a Ciência e a Tecnologia (Project POCTI/BIO/46515/2002, POCI 2010 and student grant SFRH/BD/10614/2002). vii Abstract Adenovirus vectors (AdV) have attracted considerable interest over the past decade, with ongoing clinical development programs for applications ranging from replacement therapy for protein deficiencies to cancer therapeutics to prophylactic vaccines. In fact, in the period 1989-2005, over 300 clinical trials were conducted using adenoviruses worldwide, adenovirus type 5 (Ad5) being the vector of choice. Consequently, considerable product process, analytical and formulation development has to be carried out in order to respond to the constantly increasing market requirements for AdV. One of the major restrictions on AdV production is the so-called “cell density effect”, i.e., a drop in cell specific productivity concomitant with increased cell concentration at infection (CCI) above 1106 cells/ml. This limitation on AdV production has to be fine-tuned in order to obtain higher AdV volumetric productivities; this will be the subject of study and development in this thesis. In Chapter I, a state of the art on recent developments and approaches of AdV production and purification, including successes and failures in veterinary clinical applications to date, is presented. In Chapter II, the cell density effect is explored. Thus, AdV production at cell densities higher than 1×106 cell/ml is evaluated by comparing two different serum free media (CD 293 medium from Invitrogen and EX-Cell medium from JRH). The results show that while EX-Cell medium allows the higher maximum cell number (7.5×106 cell/ml), CD 293 is the best option for AdV production at high cell densities, allowing infection at CCI of 2×106 cell/ml without significant losses in terms of cell specific productivity. xiv com compostos químicos e infectadas à mesma CCI. A maior productividade celular específica foi obtida quando as células foram sincronizadas na fase S do ciclo celular, o que confirma a nossa hipótese anterior. Assim, uma vez que o estado fisiológico das células é um parâmetro importante a considerar na altura da infecção na produção de AdV, foi desenvolvida uma estratégia para sincronizar células, sem recorrer ao uso de inibidores químicos, usando alterações da temperatura. Com esta estratégia foi possível obter uma sincronização de 57% das células na fase S do ciclo celular apenas descendo a temperatura para 31ºC durante 67h e aumentando-a novamente para 37ºC durante 72h. Ao infectar estas células sincronizadas obteve-se um aumento de 7.3 vezes na productividade celular específica. Mas, devido a perdas na viabilidade celular e aos elevados tempos de incubação necessários para obter sincronização na fase S, são necessários mais estudos para “afinar” o processo. Em anexo são mostrados dois estudos adicionais. No Anexo A, é avaliada a importância de cinéticas de infecção na optimisação do processo de produção de AdV, tais como a MOI (Multiplicidade de Infecção) e HPI (Horas Pós Infecção). Em adição, o uso de meios suplementados com soro foi comparado com um meio livre de soro em termos de crescimento celular e produção de AdV; no Anexo B, é avaliado o efeito da adição de vitaminas ou lipidos na altura da infecção a elevadas densidades celulares, sem se ter verificado um aumento relevante em termos de produção de AdV. Esta tese apresenta contribuições relevantes para a actual tecnologia de produção de AdV. Tanto a limitação do meio (quer por falta de nutrientes quer por produção excessiva de bioproductos, nomeadamente a amónia) e o estado fisiologico das células no momento da infecção foram identificados como parâmetros valiosos a ser considerados na produção de AdV a elevadas densidades celulares. xv Thesis Publications Journals Ferreira T.B., Alves P.M., Aunins J.G., Carrondo M.J.T. (2005) “Use of adenoviral vectors as veterinary vaccines”, Gene Therapy, 12:S73-S83. Ferreira T.B., Ferreira A.L., Carrondo M.J.T., Alves P.M. (2005) “Two different serum-free media and osmolality effect upon human 293 cell growth and adenovirus production”, Biotechnology Letters, 27:18091813. Ferreira T.B., Ferreira A.L., Carrondo M.J.T., Alves P.M. (2005) “Effect of refeed strategies and non-ammoniagenic medium on adenovirus production at high cell densities”, Journal of Biotechnology, 119:272280. Ferreira T.B., Carrondo M.J.T., Alves P.M. (2007) “Effect of ammonia production on intracellular pH: consequent effect on adenovirus vector production”, Journal of Biotechnology, 129:433-438. Zhang C., Ferreira T.B., Cruz P.E., Alves P.M., Haury M., Carrondo M.J.T. (2006) “The importance of 293 cell cycle phase on adenovirus vector production”, Enzyme and Microbial Technology, 39:1328-1332. Ferreira T.B., Perdigão R., Silva A.C., Zhang C., Aunins J.G., Carrondo M.J.T., Alves P.M., “293 cell cycle synchronisation in adenovirus vector production”, Biotechnology and Bioengineering (submitted). Extended Conference Proceedings Ferreira, T.B., Alves, P.M., Gonçalves, D., Carrondo, M.J.T. (2005) “Effect of MOI and medium composition on adenovirus infection xvi kinetics”, Animal Cell Technology Meets Genomics, Proceedings of the 18th ESACT Meeting, Granada, Spain, May 11-14, 2003, Edited by Gòdia, F. and Fussenegger, M., Springer, 329-332. Ferreira T.B., Alves P.M., Carrondo M.J.T. (2007) “Effect of vitamins or lipids addition on adenovirus production at high cell densities”, in Smith R (Ed), Cell Technology Processes for Cell Products, Springer (in press). Other Publications Describing Work to Which the Author has Contributed During his Thesis Work but not Included in this Thesis Peixoto C., Ferreira T.B., Carrondo M.J.T., Cruz P.E., Alves P.M. (2006) “Purification of adenoviral vectors using expanded bed chromatography”, Journal of Virological Methods, 132:121-126. Matias A., Serra A.T., Malpique R., Silva A.C., Marcelino I., Ferreira T.B., Perdigão R., Alves P., Duarte C., “Antiviral Effect of Resveratrol and a Grape Natural Extract on Adenovirus Type 5 Infection”, Journal of Agricultural and Food Chemistry (submitted). Peixoto C., Ferreira T.B., Carrondo M.J.T., Alves P.M., “New purification platform for adenovirus vectors based on membrane technology”, Biotechnology and Applied Biochemistry (submitted). xvii Contents CHAPTER I - INTRODUCTION__________________________________ 1 Use of adenoviral vectors as veterinary vaccines________________ 3 CHAPTER II - CELL DENSITY EFFECT___________________________ 43 Two different serum-free media and osmolality effect upon human 293 cell growth and adenovirus production___________________ 45 CHAPTER III - MEDIUM LIMITATIOS ____________________________ 59 Part 1 - Effect of refeed strategies and non-ammoniagenic medium on adenovirus production at high cell densities________________ 61 Part 2 - Effect of ammonia production on intracellular pH: consequent effect on adenovirus vector production______________________ 81 CHAPTER IV - CELL CYCLE _________________________________ 103 Part 1 - The importance of 293 cell cycle phase on adenovirus vector production ___________________________________________ 105 Part 2 - 293 cell cycle synchronisation in adenovirus vector production ___________________________________________ 125 CHAPTER V - DISCUSSION AND CONCLUSIONS___________________ 151 CHAPTER VI - ANNEXES____________________________________ 161 Annex A - Effect of MOI and medium composition on adenovirus infection kinetics ______________________________________ 163 Annex B - Effect of vitamins or lipids addition on adenovirus production at high cell densities __________________________ 173 CHAPTER I Introduction Introduction 3 Use of adenoviral vectors as veterinary vaccines Ferreira T.B., Alves P.M., Aunins J.G., Carrondo M.J.T. (2005) Gene Ther., 12:S73-S83. Chapter I 4 Abstract Vaccines are the most effective and inexpensive prophylactic tool in human and veterinary medicine. Ideally, vaccines should induce a lifelong protective immunity against the target pathogen while not causing clinical or pathological signs of diseases in vaccinated people or animals. However such ideal vaccines are rare in the veterinary field. Many vaccines are either of limited effectiveness or have harmful side effects. In addition, there are still severe diseases with no effective vaccines. A very important criterion for an ideal vaccine in veterinary medicine is low cost; this is especially important in developing countries and even more so for poultry vaccination, where vaccines must sell for a few cents a dose. Traditional approaches include inactivated vaccines, attenuated live vaccines and subunit vaccines. Recently, genetic engineering has been applied to design new, improved vaccines. Adenovirus vectors are highly efficient for gene transfer in a broad spectrum of cell types and species. Moreover, adenoviruses often induce humoral, mucosal and cellular immune responses to antigens encoded by the inserted foreign genes. Thus, adenoviruses have become singular vectors of choice for delivery and expression of foreign proteins for vaccination. Consequently, the market requirements for adenovirus vaccines are increasing, creating a need for production methodologies of concentrated vectors with warranted purity and efficacy. This chapter summarizes recent developments and approaches of adenovirus production and purification, including successes and failures in veterinary clinical applications to date. At the end of the chapter, a brief outline of the scope intended for this thesis is presented. Introduction 5 CONTENTS 1. ADENOVIRUS VACCINATION HISTORY _________________________ 6 2. ADENOVIRUS BIOLOGY ____________________________________ 9 3. SPECIES-SPECIFIC PLATFORMS _____________________________ 10 4. APPLICATIONS _________________________________________ 12 4.1. Adenovirus Vaccination in companion animals_____________ 12 4.1.1. Feline Immunodeficiency Virus (FIV)_____________________ 12 4.1.2. Canine Distemper Virus (CDV) __________________________ 13 4.1.3. Rabies Virus (RV)_____________________________________ 13 4.2. Adenovirus Vaccination in Poultry ______________________ 14 4.2.1. Avian Infectious Bronchitis Virus (IBV) ___________________ 14 4.2.2. Infectious Bursal Disease Virus (IBDV)____________________ 15 4.3. Adenovirus Vaccination in Swine _______________________ 15 4.3.1. Classical Swine Fever Virus (CSFV) ______________________ 15 4.3.2. Pseudorabies Virus (PrV)_______________________________ 16 4.3.3. Foot and Mouth Disease Virus (FMDV)____________________ 17 4.3.4. Porcine Respiratory and Reproductive Syndrome Virus (PRRSV)19 4.3.5. Transmissible Gastroenteritis (Corona) Virus (TGEV) _______ 19 4.3.6. Swine Influenza Virus (SIV)_____________________________ 20 4.4. Adenovirus Vaccination in Cattle _______________________ 21 4.4.1. Bovine Viral Diarrhea Virus (BVDV) ______________________ 21 4.4.2. Bovine Parainfluenza Virus Type 3 (bPIV3)________________ 22 4.4.3. Bovine Herpes Virus 1 (BHV-1)__________________________ 23 4.4.4. Rinderpest Virus (RPV) and Peste des Petits Ruminants Virus (PPRV) ___________________________________________________ 24 5. ADENOVIRUS PRODUCTION AND PURIFICATION AS VETERINARY VACCINES _______________________________________________ 26 6. SCOPE OF THE THESIS ___________________________________ 30 7. ACKNOWLEDGEMENTS ___________________________________ 31 8. REFERENCES___________________________________________ 32 Chapter I 12 with BAd3 failed to produce either clinical signs or gross lesions, but all animals seroconverted. The cell entry pathway of BAd3 is not well understood (Benko and Harrach 1998; Reddy et al., 2000). 4. APPLICATIONS 4.1. Adenovirus Vaccination in companion animals 4.1.1. Feline Immunodeficiency Virus (FIV) The infection of cats with FIV results in an immunosuppressive disease that is transmitted by blood and saliva. FIV invades and destroys the monocyte/macrophage system and infects B-cells (Uhl et al., 2002). Different approaches have been made trying to immunize cats against FIV; so far, no effective vaccine has been made. Gonin et al., (1995), 10 years ago, constructed a replication-defective HAd5, containing the envelope protein ENV gene of FIV; however, despite the fact that an antibody response to pseudorabies virus in cats showed the potential of rHAd5 vectors to be used in this species, cats injected with 1010.8-1011.8 of 50 percent tissue culture infectious dose (TCID50) adjuvanted with montanide ISA 708 (water in non mineral oil) or with montanide ISA 206 (double water/mineral oil/water) of this AdV did not develop detectable antibody response against ENV. Moreover, it was observed that even if high titers of antibodies against ENV products are induced, they could still be insufficient for protection (Gonin et al., 1995). Since then, no further work has apparently been performed concerning the use of AdV as a potential vaccine against FIV. Introduction 13 4.1.2. Canine Distemper Virus (CDV) CDV induces fatal diseases including encephalitis with demyelination, diarrhea and respiratory disorders in dogs. Although conventional live modified vaccines are commercially available and widely used in the field, their efficacy is limited in the presence of maternally-derived antibodies (Barrett 1999). Thus, a new and improved CDV vaccine which could overcome this limitation would constitute a significant improvement. The construction and characterization of the first replication-competent rCAd2 started recently. The genes that code for CDV hemagglutinin (HA) (Fischer et al., 2002; Hirama et al., 2003) or fusion (F) (Fischer et al., 2002) proteins were inserted in two CAd2s and used as a candidate vaccine in puppies. It was reported that intranasal vaccination with a mixture containing 105.8 TCID50 of which rCAd2 provided an excellent level of protection in seronegative puppies, inducing almost complete protection. In contrast, intranasal immunization of puppies born to CDV and CAd2 immune dams failed to activate specific and protective immune responses. However, when the same puppies were vaccinated subcutaneously, significant seroconversion and solid protective immunity were triggered. Furthermore, a significant priming of memory responses was evidenced immediately after challenge, this constituting an efficient strategy to overcome both passive and active Ad specific immunity in the dog. 4.1.3. Rabies Virus (RV) Rabies still presents a health threat not only to humans but also to dogs and cats, being the dog the only important vector for humans, especially in less developed nations where uncontrolled canine rabies often is endemic. A replication-competent and -defective HAd5 expressing a rabies glycoprotein (RG) has been developed, inducing immunity to rabies in rodent, canine, foxes and skunk when given by intramuscular, Chapter I 14 subcutaneous or intranasal routes with a dose of 108 TCID50/animal. However, oral immunization failed to induce a measurable antibody response to RV using the same dose (Xiang et al., 1996; Vos et al., 2001). Dogs previously vaccinated with commercially available vaccines, immunized with 107 plaque forming unit (pfu)/animal of a replicationdefective HAd5 expressing the RG, have developed higher titers of viral neutralizing antibodies against RV 10 days after vaccination when compared with conventional vaccines under similar conditions. Moreover, the immunization of dogs with the commercial non Ad vaccines is required yearly or, at best, every 3 years, depending on the type of vaccine. On the other hand, the higher antibody titers obtained with Ad vaccines against RV would reduce the frequency of dog immunization, reducing the costs for pet owners (Tims et al., 2000). One important advantage of this recombinant vaccine is the fact that the immune response to the RG was shown not to be impaired by maternal immunity; thus, this Ad vaccine is highly suitable for neonatal immunization (Wang et al., 1997). 4.2. Adenovirus Vaccination in Poultry 4.2.1. Avian Infectious Bronchitis Virus (IBV) IBV is a highly contagious pathogen of poultry causing significant morbidity and mortality. Depending on the strain, IBV can target the respiratory tract, kidney and oviducts and result in nephritis and reduced egg production. In addition, more than 20 IBV serotypes have been identified worldwide and new serotypic variants have been identified as a result of the widespread use of live attenuated vaccines. Different approaches have been developed in order to generate a more efficacious vaccine against IBV. The expression of S1, a glycoprotein involved in the attachment of cellular receptors, by a vaccinia virus was able to induce virus-neutralizing antibodies to IBV when delivered to mice; however, multiple injections are Introduction 15 required to achieve a reasonable degree of protection. Recently, a FAd expressing the S1 of IBV has been developed (Johnson et al., 2003). A single dose of 106 TCID50/animal was shown to be sufficient to obtain complete protection of chickens at the trachea, the primary site of infection by IBV. Moreover, even in the face of FAV maternal antibodies, a high level of protection was achieved (Johnson et al., 2003). 4.2.2. Infectious Bursal Disease Virus (IBDV) IBDV induces an immunosuppressive disease of chickens by destruction of the B-lymphocytes. Current vaccination alternatives consist of either live virus or inactivated oil-emulsion vaccines, which induce serum antibody production in breeding hens after natural exposure; they are transferred to the progeny chicks via the yolk sac providing protection for the first critical weeks after hatching (Muller et al., 2003). The construction of a rFAd10 containing the VP2 gene from IBDV has been described (Sheppard et al., 1998). This recombinant vaccine was shown to induce an immune response in chickens to VP2 after vaccination with 107 pfu/animal. Moreover, after challenge with IBDV, intravenously, intraperitoneally, subcutaneously or intramuscularly vaccinated chickens were protected, although no protection was observed in conjunctival sac vaccinated chickens (Sheppard et al., 1998). 4.3. Adenovirus Vaccination in Swine 4.3.1. Classical Swine Fever Virus (CSFV) Classical swine fever (CSF), also known as hog cholera, is a serious and contagious viral disease of pigs with a high mortality rate, difficult to control in areas of high pig or wild boar densities, being the most economically important disease of swine in areas of intensive pig farming. Chapter I 16 For this reason, CSF is included in the A list of infectious diseases of the highest importance for international trade. Thus, it is highly relevant to develop efficacious vaccines for the control of CSFV in domestic pigs and in wild boar. Prophylactic vaccination is still carried out in many parts of the world (Paton and Greiser-Wilke 2003). New vaccine developments have included a number of different strategies for delivering the major envelope glycoprotein, E2, against which most neutralizing antibodies are directed (van Oirschot 2003). Hammond et al. (2000), constructed the first rPAd expressing the E2 protein and showed that a single dose of 107 TCID50/animal in tissue culture supernatant, when administered subcutaneously, is sufficient to completely protect pigs against subcutaneously challenge. Later, the same authors showed that when the challenge is administered orally, only 60% of the animals were protected (Hammond et al., 2001b); more, recently, it was shown that pigs given two oral doses of 106 TCID50/animal of rPAd were completely protected from the disease (Hammond et al., 2003). 4.3.2. Pseudorabies Virus (PrV) PrV is an alpha herpesvirus which causes the economically important and widespread Aujeszky’s disease (AJD) in pigs. PrV is a highly neurotropic virus causing nervous and respiratory complications in pigs, the natural host, and in a variety of other animal species. Vaccination against AJD is widely practised with live attenuated or killed whole virus vaccines. However, neonatal immunization is often limited in the presence of maternally-derived antibody, which inhibits the immune response against both vaccines (Roth 1999). Recently, the use of rAd vaccines carrying individual PrV genes were constructed as a safe alternative. Glycoproteins gD, gB and gC of PrV were chosen on the basis of their role in eliciting a protective immune response against virus infection. It was shown that piglets vaccinated intramuscularly with 108.6-109.6 TCID50/animal one day Introduction 17 after birth, with replication-competent HAd5 harbouring these three genes, developed similar neutralizing antibody responses independently of the presence or absence of maternal antibodies and were partially protected against challenge 16 weeks later (Monteil et al., 2000). For pigs that are slaughtered a few months after birth, one-shot vaccination at birth could provide protection of sufficient duration. Finally, two doses of 2 ml of clarified tissue culture supernatant containing 105.4 TCID50 of rPAd expressing the PrV gD gene were administered subcutaneously and showed to protect pigs after challenge. Post-morten, gross lesions of pneumonia were found in the lungs of pigs given a single dose of vaccine; the lungs of pigs given two doses were free from disease (Hammond et al., 2001a). 4.3.3. Foot and Mouth Disease Virus (FMDV) Foot and mouth disease (FMD) is a severe, clinically acute, vesicular disease of cloven-hoofed animals including domesticated ruminants, pigs and more than 70 wildlife species. Pigs are recognized as a significant factor in the spread of the disease since a single pig releases as much aerosol virus as 3000 cattle in a short period of time (Alexandersen et al., 2003; Thomson et al., 2003). The economic and social impact of FMD can be catastrophic when an outbreak occurs in FMD-free countries populated with immunologically naive animals. The current vaccine is a chemically inactivated preparation of concentrated infected cell culture supernatant. However, FMD-free countries generally prohibit its use because of the lack of an approved diagnostic test that can reliably distinguish vaccinated from infected animals. Moreover, current vaccines can induce a protective response only after approximately 7 days post vaccination; this is a critical issue in disease-free countries, where, in the case of FMD outbreaks, a rapid control in preventing the spread of the disease is crucial (see Doel (2003) for a review). Thus, there is a need to develop disease control strategies relying on more rapidly induced protection. Mayr et al. (1999) Chapter I 18 developed a replication-defective HAd5 vector containing the capsid polypeptide P1 and the viral 3C protease coding regions, necessary for processing P1 to the capsid proteins VP0, VP3 and VP1, from the FMDV strain A12. They observed that vaccinated swine with 1×108 pfu/animal in PBS developed antibodies against FMDV structural proteins and an FMDVspecific neutralizing antibody response which seems to increase slightly by boosting the swine with a second inoculation of 5×108 pfu/animal in PBS at 4 weeks post-initial vaccination (Mayr et al., 2001). This vaccination protocol was shown to offer a significant degree of protection to the pigs, as five of six pigs were completely protected, while the remaining animal had significantly reduced signs of disease. Later, Moraes et al. (2002) showed that a single dose of 5×109 pfu/animal in PBS of a replicationdefective HAd5 expressing the P1 coding region of FMDV strain A24 completely protected pigs against homologous challenge 7, 14 or 42 days after vaccination. Since efficacious vaccination is strain dependent and the infection with one serotype does not confer protection against another, HAd5 bicistronic vector vaccines have been developed (Wu et al., 2003), decreasing the cost for multivalent adenoviral FMD vaccines. For the construction of these vectors, the P1 capsid coding region for both A24 and O1 strains and the 3C protease coding region of A12 strain were used. However, the neutralizing antibody response after vaccination with 2.5×109 pfu/animal in PBS of the bicistronic vector was considerably lower than that induced by a commercial FMD vaccine or the monovalent Ad-A24 vaccine (Moraes et al., 2002). Recently, a new strategy based on the fact that FMDV is highly sensitive to alpha/beta interferon (IFN-/) have been developed using a replication-defective HAd5 (Chinsangaram et al., 2003; Moraes et al., 2003). This strategy has been shown to completely protect pigs after vaccination with 109 pfu/animal when challenged 24 h later with virulent FMDV. Introduction 19 4.3.4. Porcine Respiratory and Reproductive Syndrome Virus (PRRSV) PRRSV is the causative agent of an economically important pig disease, with a worldwide distribution, characterized by reproductive failure in sows and respiratory problems in unweaned and growing pigs. Swine macrophage is the only cell type known to support PRRSV replication, making commercial production impossible. Direct contact between infected and naive pigs is the predominant route of PRRSV transmission. Moreover, pneumonia caused by PRRSV infection is more severe in young pigs compared to adults and may be complicated by concurrent bacterial infection (Rossow 1998). Gonin et al. (1999) observed that PRRSV infected pigs present circulating antibodies responsible for viral neutralization mainly directed against GP5, an envelope protein. Gagnon et al. (2003) constructed a replication-defective HAd5 expressing the GP5 protein and used this recombinant virus to immunize pigs using two intradermal injections with 5×108-1×109pfu/animal in a mixture containing 100 L of PBS and 100 L of poloxamer SP1017 at 0.02%. It was observed that following challenge given intranasally 14 days after the booster, pigs produced high antibody titers to GP5 protein. Moreover, vaccinated pigs presented specific immune memory which, following a subsequent PRRSV infection, resulted in a rapid clonal expansion of memory cells to the neutralizing epitopes of the authentic viral GP5 protein. 4.3.5. Transmissible Gastroenteritis (Corona) Virus (TGEV) TGEV infects the enteric and respiratory tissues of newborn piglets resulting in mortalities approaching 100%. The virus infects epithelial cells and, in some cases, lung macrophages (Garwes 1988). There are several commercially available TGEV vaccines, both inactivated and attenuated; these do not fully protect piglets. Several attempts have been made to Chapter I 20 develop efficacious recombinant TGEV vaccines (Tuboly and Nagy 2001). The spike protein (S) was identified as the major inducer of TGEVneutralizing antibodies and it mediates binding of TGEV to its cellular receptor (Jimenez et al., 1986). Thus, HAd5 expressing the S protein was constructed and used to study the induction of antibodies providing protection in swine (Torres et al., 1995). It was observed that porcine serum, elicited by 109 pfu/animal in PBS of this recombinant, when mixed with a lethal dose of virus prior to administration to susceptible pigs, prevented the replication of virulent TGEV administered orally as virusantibody mixtures and fully protected swine from clinical signs and death. Moreover, the used dose did not produce any clinical symptoms in any of the more than 50 animals inoculated up to 10 weeks after inoculation (Torres et al., 1996), suggesting that this vector can be used as a live vaccine in swine without secondary complications associated with the vector. However, a PAd vector could be more effective than HAd. Thus, Tuboly and Nagy (2001) constructed a rPAd5 expressing the TGEV protein. It was observed that a single oral dose of 5×106 pfu/animal of the recombinant virus was sufficient to induce both a systemic and a local humoral immune response (Tuboly and Nagy 2001). Unfortunately challenge experiments were not carried out. 4.3.6. Swine Influenza Virus (SIV) SIV is a widespread and important pathogen in species as diverse as poultry, swine, marine mammals and humans. In pigs, influenza can occur either as an enzootic problem in a herd or, more commonly, as explosive outbreaks of acute respiratory disease. Although rarely fatal, swine influenza can be of substantial economic impact (van Reeth and Nauwynck 2000). In addition, there is growing concern for the potential for synergistic infections with influenza and PRRSV. Beyond the impact of influenza for the swine industry, pigs are also very important in the global ecology of Introduction 21 influenza A viruses in humans (Zhou et al., 1999). SIV vaccines that are commercially-available are inactivated, whole-virus or subunit vaccines. While these vaccines may decrease the incidence and severity of clinical disease, they do not consistently provide complete protection from virus infection. Two replication-defective HAd5 were developed as potential vaccines against SIV: rHAd5 expressing the influenza virus H3 haemagglutinin (HA) (Tang et al., 2002; Wesley et al., 2004) inducing predominately a subtype-specific humural immune response (Macklin et al., 1998); and rHAd5 expressing the nucleoprotein (NP) (Wesley et al., 2004), a group-specific stimulating cytotoxic T lymphocytes for cross-reactive immunity to all influenza A subtypes (Ulmer et al., 1998). It was observed that the immunization of mice with 5×108 TCID50/animal and pigs with 2×1010 TCID50/animal in PBS with the first recombinant, described above, developed high levels of virus-specific hemagglutination-inhibition antibody to SIV by 4 weeks post vaccination and the animals were partially protected. On the other hand, pigs vaccinated with 2×1010 TCID50/ml in PBS of both recombinant viruses in a mixture were completely protected. 4.4. Adenovirus Vaccination in Cattle 4.4.1. Bovine Viral Diarrhea Virus (BVDV) BVDV is responsible for reduced milk production, reduced reproductive performance and growth retardation. In addition, acute infection in adult cattle, congenital defects and increased neonatal mortality are also clinical manifestations of BVDV infection (Goens 2002). Finally, it was reported that BVDV may play an indirect role in immunosuppression (Wellenberg et al., 2002). Currently, inactivated and modified-live vaccines are used; however, both types of vaccines have significant shortcomings. A rHAd5 expressing the nucleocapsid C protein (p14), that is highly conserved among many different pestiviruses, to which BVDV belongs, was Chapter I 28 batch mode providing the easiest way to proceed as no extra feeding is required and the risk of contamination is lowered given the simplicity of operation; ii) fed-batch mode, easy to operate and readily scalable, is employed to extend culture lifetime by supplementing limiting nutrients or reducing the accumulation of toxic metabolites; and iii) perfusion mode, consisting in cell retention at a relatively high concentration inside the bioreactor, while fresh nutrient supply and metabolite removal takes place (see Kamen and Henry (2004) for a review). Methodologies for production of concentrated AdV at low cost are mandatory as the market needs for Ad are increasing. Although Ad has the advantage to be produced at high titers (1010-1011 pfu/ml) (Babiuk and Tikoo 2000), to obtain a good immune response in a large proportion of treated animals, particularly for mucosal vaccination, large doses and thus culture volumes are required. Further process development aiming at higher yields of product is clearly necessary. Improvements in volumetric production can be achieved by increasing the cell density at which cells can be infected without lowering the specific yield of the product; however, production of AdV that maintain a high specific yield in batch operation is limited to cell densities in the range of 1×106 cell/ml; several approaches have been made in order to overcome this so called “cell density effect” (Nadeau and Kamen 2003): Garnier et al. (1994) demonstrated that medium replacement at infection and the addition of glucose at 24 hours post infection (hpi) together with periodical pH adjustments, allowed a sustained maximum specific productivity at 1.6106 cell/ml whereas Nadeau et al. (1996), further improved this strategy by also adding essential amino acids at 24 hpi thereby stabilising volumetric productivity at cell densities above 2106 and below 3106 cell/ml. Such results hint at the existence of substrate limitation and/or byproduct inhibition at high cell densities. At production scale, medium exchange will increase the cost of the final product, which is even more critical when this is to be used in the veterinary field. Perfusion mode operation has been attempted as a Introduction 29 means to control the culture environment and remove toxic byproducts (Henry et al., 2004; Kamen and Henry 2004); nevertheless, the cell specific productivity could only be maintained by infecting cells at densities up to 3106 cell/ml using high perfusions rates of 2 reactor volumes per day, at 2 days post infection, a very costly proposition (Kamen and Henry 2004). The importance of infection kinetics on Ad production and the significance of variables such as Multiplicity Of Infection (MOI) and harvesting time in process optimisation is also mandatory to increase production yields, to avoid rapid depletion of costly and certified master virus banks as well as to ensure that the infection kinetics is reproducible between different production scales. Moreover, the use of low MOI at large scale would be preferential since an intermediate step of virus inoculum production would be avoided (Annex A). Finally, veterinary products should be purified with a minimal number of steps and the unit operations employed should be simple and nonexpensive. Traditionally, laboratory purification of rAd was achieved using two rounds of cesium chloride (CsCl) density gradient ultracentrifugation. However, CsCl density gradient method is not scaleable. Despite the fact that chromatographic purification is an expensive method, ion exchange, hydrophobic interaction, metal chelate and size-exclusion chromatography have been evaluated for capture and purification of HAd5 (Huyghe et al., 1995). Kamen and Henry (2004) developed a method for gene therapy application consisting of: i) harvest of infected cells by continuous centrifugation, ii) cell lysis by osmotic shock, iii) DNAse treatment with centrifugal/conditioning, iv) filtration, v) anion-exhange chromatography, vi) ultrafiltration/concentration and vii) size exclusion chromatography. This protocol allows large scale purification of AdV with purity comparable to the CsCl gradient method; however, it increases the final product price, well beyond the reach of veterinary utilisation. Introgen developed a method consisting of a single ion chromatography run, after concentration/diafiltration and nuclease treatment, with a total recovery Chapter I 30 of the virus product of 70% (Zhang et al., 2001). Also, a protocol to purify HAd5 from the bulk harvested directly from the bioreactor after lysis (without a concentration step) was developed using a single ion exchange chromatographic step followed by ultrafiltration with a final yield of 32% achieved in less than one working day with a minimal amount of sample manipulation (Peixoto et al., 2005). This process presented the advantage of capturing the viral particles directly from cellular extracts and to account for the significant amounts (25-80%) of AdV that are present in the culture medium at harvest time due to early cell lysis (Schagen et al., 2000). 6. SCOPE OF THE THESIS One of the mains goals of a commercial AdV cultivation process for gene therapy or vaccine applications is to maximise the virus volumetric productivity, defined as units of viral infective particles, per culture volume per unit time. Since one can infect at different cell concentrations, the cell specific productivity, defined as viral infective particles produced per cell and unit time, is a parallel concept; ideally, operating at the highest possible cell specific productivity obtained at the highest cell concentration at infection would maximize the volumetric productivity of the process but quality of the produced AdV is of paramount importance. This thesis has attempted to improve the current manufacturing process of AdV production by performing infection at high cell densities maintaining the cell specific productivity. To achieve this, the bottlenecks in cell productivity have been studied as well as the parameters that may affect the vector production at high cell densities. The cell density at infection is a very important parameter as it strongly impacts the AdV volumetric productivity. However, in simple batch-mode operations, a relatively narrow range for optimal cell density at infection of 1×106 cell/ml has been consistently reported. The reasons for the drop in per cell productivity at Introduction 31 higher cell densities are not currently known, as medium formulations allow maximum cell growths reaching 8×106 cell/ml. Therefore more research is needed to understand and overcome this phenomenon to enable simple batch processes to reach higher productivities. In order to design a rational production process that allow the maximisation of the AdV yields at high cell densities, four different bioreaction approaches were explored in this thesis: i) media exchange, ii) refeed strategy, iii) use of nonammoniagenic medium and iv) cell cycle synchronisation. The knowledge gained in this scientific excursion is integrated into an improved process. 7. ACKNOWLEDGEMENTS The authors acknowledge and appreciate the financial support received from the European Commission (Project RP/PPR ORALVAC ICA4-CT-200030027) and from Fundação para a Ciência e Tecnologia – Portugal (Project POCTI/BIO/46515/2002) and student grant (SFRH/BD/10614/2002)). Chapter I 32 8. REFERENCES Alexandersen, S., Zhang, Z., Donaldson, A.I. and Garland, A.J., 2003. The pathogenesis and diagnosis of foot-and-mouth disease. J Comp Pathol 129, 1-36. Babiuk, L.A. and Tikoo, S.K., 2000. Adenoviruses as vectors for delivering vaccines to mucosal surfaces. J Biotechnol 83, 105-13. Babiuk, L.A., van Drunen Littel-van den Hurk, S., Tikoo, S.K., Lewis, P.J. and Liang, X., 1996. Novel viral vaccines for livestock. Vet Immunol Immunopathol 54, 355-63. Barrett, T., 1999. Morbillivirus infections, with special emphasis on morbilliviruses of carnivores. Vet Microbiol 69, 3-13. Baxi, M.K., Deregt, D., Robertson, J., Babiuk, L.A., Schlapp, T. and Tikoo, S.K., 2000. Recombinant bovine adenovirus type 3 expressing bovine viral diarrhea virus glycoprotein E2 induces an immune response in cotton rats. Virology 278, 234-43. Benko, M. and Harrach, B., 1998. A proposal for a new (third) genus within the family Adenoviridae. Arch Virol 143, 829-37. Bett, A.J., Prevec, L. and Graham, F.L., 1993. Packaging capacity and stability of human adenovirus type 5 vectors. J Virol 67, 5911-21. Bloom, B.R. and Lambert, P. 2003. The vaccine book. Press, A., editor. California. Breker-Klassen, M.M., Yoo, D., Mittal, S.K., Sorden, S.D., Haines, D.M. and Babiuk, L.A., 1995. Recombinant type 5 adenoviruses expressing bovine parainfluenza virus type 3 glycoproteins protect Sigmodon hispidus cotton rats from bovine parainfluenza virus type 3 infection. J Virol 69, 4308-15. Chinsangaram, J., Moraes, M.P., Koster, M. and Grubman, M.J., 2003. Novel viral disease control strategy: adenovirus expressing alpha interferon rapidly protects swine from foot-and-mouth disease. J Virol 77, 1621-5. Danthinne, X. and Imperiale, M.J., 2000. Production of first generation adenovirus vectors: a review. Gene Ther 7, 1707-14. Introduction 33 Diallo, A., 2003. Control of peste des petits ruminants: classical and new generation vaccines. Dev Biol (Basel) 114, 113-9. Doel, T.R., 2003. FMD vaccines. Virus Res 91, 81-99. Dunham, S.P., 2002. The application of nucleic acid vaccines in veterinary medicine. Res Vet Sci 73, 9-16. Elahi, S.M., Shen, S.H., Talbot, B.G., Massie, B., Harpin, S. and Elazhary, Y., 1999a. Induction of humoral and cellular immune responses against the nucleocapsid of bovine viral diarrhea virus by an adenovirus vector with an inducible promoter. Virology 261, 1-7. Elahi, S.M., Shen, S.H., Talbot, B.G., Massie, B., Harpin, S. and Elazhary, Y., 1999b. Recombinant adenoviruses expressing the E2 protein of bovine viral diarrhea virus induce humoral and cellular immune responses. FEMS Microbiol Lett 177, 15966. Fischer, L., Tronel, J.P., Pardo-David, C., Tanner, P., Colombet, G., Minke, J. and Audonnet, J.C., 2002. Vaccination of puppies born to immune dams with a canine adenovirus-based vaccine protects against a canine distemper virus challenge. Vaccine 20, 3485-97. Freshney, R.I. 1986. Animal cell culture: A pratical approach. Rickwood, D. and Hames, B.D., editors: IRL Press. 248 p. Gagnon, C.A., Lachapelle, G., Langelier, Y., Massie, B. and Dea, S., 2003. Adenoviral-expressed GP5 of porcine respiratory and reproductive syndrome virus differs in its cellular maturation from the authentic viral protein but maintains known biological functions. Arch Virol 148, 951-72. Garnier, A., Cote, J., Nadeau, I., Kamen, A. and Massie, B., 1994. Scale-up of the adenovirus expression system for the production of recombinant protein in human 293S cells. Cytotechnology 15, 145-155. Garwes, D.J., 1988. Transmissible gastroenteritis. Vet Rec 122, 462-3. Chapter I 34 Giese, M., 1998. DNA-antiviral vaccines: new developments and approaches--a review. Virus Genes 17, 219-32. Goens, S.D., 2002. The evolution of bovine viral diarrhea: a review. Can Vet J 43, 946-54. Gogev, S., Vanderheijden, N., Lemaire, M., Schynts, F., D'Offay, J., Deprez, I., Adam, M., Eloit, M. and Thiry, E., 2002. Induction of protective immunity to bovine herpesvirus type 1 in cattle by intranasal administration of replication-defective human adenovirus type 5 expressing glycoprotein gC or gD. Vaccine 20, 1451-65. Gonin, P., Fournier, A., Oualikene, W., Moraillon, A. and Eloit, M., 1995. Immunization trial of cats with a replication-defective adenovirus type 5 expressing the ENV gene of feline immunodeficiency virus. Vet Microbiol 45, 393-401. Gonin, P., Pirzadeh, B., Gagnon, C.A. and Dea, S., 1999. Seroneutralization of porcine reproductive and respiratory syndrome virus correlates with antibody response to the GP5 major envelope glycoprotein. J Vet Diagn Invest 11, 20-6. Graham, F.L., Smiley, J., Russell, W.C. and Nairn, R., 1977. Characteristics of a human cell line transformed by DNA from human adenovirus type 5. J Gen Virol 36, 59-74. Hammond, J.M., Jansen, E.S., Morrissy, C.J., Hodgson, A.L. and Johnson, M.A., 2003. Protection of pigs against 'in contact' challenge with classical swine fever following oral or subcutaneous vaccination with a recombinant porcine adenovirus. Virus Res 97, 151-7. Hammond, J.M., Jansen, E.S., Morrissy, C.J., van der Heide, B., Goff, W.V., Williamson, M.M., Hooper, P.T., Babiuk, L.A., Tikoo, S.K. and Johnson, M.A., 2001a. Vaccination of pigs with a recombinant porcine adenovirus expressing the gD gene from pseudorabies virus. Vaccine 19, 3752-8. Hammond, J.M., Jansen, E.S., Morrissy, C.J., Williamson, M.M., Hodgson, A.L. and Johnson, M.A., 2001b. Oral and sub-cutaneous vaccination of commercial pigs with a recombinant porcine adenovirus expressing the classical swine fever virus gp55 gene. Arch Virol 146, 1787-93. Introduction 35 Hammond, J.M., McCoy, R.J., Jansen, E.S., Morrissy, C.J., Hodgson, A.L. and Johnson, M.A., 2000. Vaccination with a single dose of a recombinant porcine adenovirus expressing the classical swine fever virus gp55 (E2) gene protects pigs against classical swine fever. Vaccine 18, 1040-50. Harui, A., Suzuki, S., Kochanek, S. and Mitani, K., 1999. Frequency and stability of chromosomal integration of adenovirus vectors. J Virol 73, 6141-6. Henry, O., Dormond, E., Perrier, M. and Kamen, A., 2004. Insights into adenoviral vector production kinetics in acoustic filter-based perfusion cultures. Biotechnol Bioeng 86, 765-774. Hirama, K., Togashi, K., Wakasa, C., Yoneda, M., Nishi, T., Endo, Y., Miura, R., Tsukiyama-Kohara, K. and Kai, C., 2003. Cytotoxic T-lymphocyte activity specific for hemagglutinin (H) protein of canine distemper virus in dogs. J Vet Med Sci 65, 109-12. Huyghe, B.G., Liu, X., Sutjipto, S., Sugarman, B.J., Horn, M.T., Shepard, H.M., Scandella, C.J. and Shabram, P., 1995. Purification of a type 5 recombinant adenovirus encoding human p53 by column chromatography. Hum gene ther 6, 1403-1416. Imler, J.-L., 1995. Adenovirus vectors as recombinant viral vaccines. Vaccine 13, 1143-1151. Jalava, K., Hensel, A., Szostak, M., Resch, S. and Lubitz, W., 2002. Bacterial ghosts as vaccine candidates for veterinary applications. J Control Release 85, 17-25. Jimenez, G., Correa, I., Melgosa, M.P., Bullido, M.J. and Enjuanes, L., 1986. Critical epitopes in transmissible gastroenteritis virus neutralization. J Virol 60, 131-9. Johnson, M.A., Pooley, C., Ignjatovic, J. and Tyack, S.G., 2003. A recombinant fowl adenovirus expressing the S1 gene of infectious bronchitis virus protects against challenge with infectious bronchitis virus. Vaccine 21, 2730-6. Chapter I 36 Jooss, K., Ertl, H.C. and Wilson, J.M., 1998. Cytotoxic T-lymphocyte target proteins and their major histocompatibility complex class I restriction in response to adenovirus vectors delivered to mouse liver. J Virol 72, 2945-54. Kamen, A. and Henry, O., 2004. Development and optimization of an adenovirus production process. J Gene Med 6 Suppl 1, S184-192. Lewis, A.M., Jr., 1998. SV40 in adenovirus vaccines and adenovirus-SV40 recombinants. Dev Biol Stand 94, 207-16. Macklin, M.D., McCabe, D., McGregor, M.W., Neumann, V., Meyer, T., Callan, R., Hinshaw, V.S. and Swain, W.F., 1998. Immunization of pigs with a particlemediated DNA vaccine to influenza A virus protects against challenge with homologous virus. J Virol 72, 1491-6. Mayr, G.A., Chinsangaram, J. and Grubman, M.J., 1999. Development of replication-defective adenovirus serotype 5 containing the capsid and 3C protease coding regions of foot-and-mouth disease virus as a vaccine candidate. Virology 263, 496-506. Mayr, G.A., O'Donnell, V., Chinsangaram, J., Mason, P.W. and Grubman, M.J., 2001. Immune responses and protection against foot-and-mouth disease virus (FMDV) challenge in swine vaccinated with adenovirus-FMDV constructs. Vaccine 19, 2152-62. Miller, E., Goldacre, M., Pugh, S., Colville, A., Farrington, P., Flower, A., Nash, J., MacFarlane, L. and Tettmar, R., 1993. Risk of aseptic meningitis after measles, mumps, and rubella vaccine in UK children. Lancet 341, 979-82. Mittal, S.K., Papp, Z., Tikoo, S.K., Baca-Estrada, M.E., Yoo, D., Benko, M. and Babiuk, L.A., 1996. Induction of systemic and mucosal immune responses in cotton rats immunized with human adenovirus type 5 recombinants expressing the full and truncated forms of bovine herpesvirus type 1 glycoprotein gD. Virology 222, 299309. Monteil, M., Le Pottier, M.F., Ristov, A.A., Cariolet, R., L'Hospitalier, R., Klonjkowski, B. and Eloit, M., 2000. Single inoculation of replication-defective Introduction 37 adenovirus-vectored vaccines at birth in piglets with maternal antibodies induces high level of antibodies and protection against pseudorabies. Vaccine 18, 1738-42. Moraes, M.P., Chinsangaram, J., Brum, M.C. and Grubman, M.J., 2003. Immediate protection of swine from foot-and-mouth disease: a combination of adenoviruses expressing interferon alpha and a foot-and-mouth disease virus subunit vaccine. Vaccine 22, 268-79. Moraes, M.P., Mayr, G.A., Mason, P.W. and Grubman, M.J., 2002. Early protection against homologous challenge after a single dose of replication-defective human adenovirus type 5 expressing capsid proteins of foot-and-mouth disease virus (FMDV) strain A24. Vaccine 20, 1631-9. Muller, H., Islam, M.R. and Raue, R., 2003. Research on infectious bursal disease-- the past, the present and the future. Vet Microbiol 97, 153-65. Murakami, P., Pungor, E., Files, J., Do, L., van Rijnsoever, R., Vogels, R., Bout, A. and McCaman, M., 2002. A single short stretch of homology between adenoviral vector and packaging cell line can give rise to cytopathic effect-inducing, helperdependent E1-positive particles. Hum Gene Ther 13, 909-20. Mutwiri, G., Bateman, C., Baca-Estrada, M.E., Snider, M. and Griebel, P., 2000. Induction of immune responses in newborn lambs following enteric immunization with a human adenovirus vaccine vector. Vaccine 19, 1284-93. Nadeau, I., Garnier, A., Côté, J., Massie, B., Chavarie, C. and Kamen, A., 1996. Improvement of recombinant protein production with human adenovirus/293S expression system using fed-batch Strategies. Biotechnol Bioeng 51, 613-623. Nadeau, I. and Kamen, A., 2003. Production of adenovirus vector for gene therapy. Biotechnol Adv 20, 475-489. Paoletti, E., 1996. Applications of pox virus vectors to vaccination: an update. Proc Natl Acad Sci U S A 93, 11349-53. Papp, Z., Babiuk, L.A. and Baca-Estrada, M.E., 1999. The effect of pre-existing adenovirus-specific immunity on immune responses induced by recombinant Cell Density Effect 45 Two different serum-free media and osmolality effect upon human 293 cell growth and adenovirus production Ferreira T.B., Ferreira A.L., Carrondo M.J.T., Alves P.M. (2005) Biotechnol. Lett., 27:1809-1813. Chapter II 46 Abstract Adenoviruses are promising vectors for gene therapy and vaccination protocols. Consequently, the market requirements for adenovirus are increasing, driving the search for new methodologies for large-scale production of concentrated vectors with warranted purity and efficacy, in a cost-effective way. Nevertheless, the production of adenovirus is currently limited by the so-called “cell density effect”, i.e., a drop in cell specific productivity concomitant with increased cell concentration at infection. In this study, two different serum-free culture media (CD 293 and EX-Cell) are evaluated in their effect on human 293 cells growth and adenovirus production at cell densities higher than 1106 cells/ml. The results show that EX-Cell was the better medium for cell growth. Although adenovirus production was equivalent in both CD 293 and EX-Cell media when the infection was performed at 1106 cells/ml, at 3106 cells/ml CD 293 proved to be the better medium. This result was related with the high ammonia content in EX-Cell medium at the highest cell concentration at infection. Besides this, the large-scale production of these vectors at high cell densities often requires refeed strategies, which increase medium osmolality. While a negative effect on cell growth was observed with increasing osmolalities, adenovirus productivity was only affected for osmolalities higher than 430 mOsm. Cell Density Effect 47 CONTENTS 1. INTRODUCTION_________________________________________ 48 2. MATERIALS AND METHODS ________________________________ 49 2.1. Cell line and culture conditions ________________________ 49 2.2. Infection in shake flasks ______________________________ 49 2.3. Virus culture samples preparation ______________________ 50 2.4. Analytical Methods __________________________________ 50 3. RESULTS AND DISCUSSION ________________________________ 50 3.1. Effect of media on cell growth _________________________ 50 3.2. Effect of media on AdV productivity_____________________ 51 3.3. Effect of osmolality increase on cell growth and AdV productivity _____________________________________________________ 53 4. ACKNOWLEDGEMENT ____________________________________ 56 5. REFERENCES___________________________________________ 57 Chapter II 48 1. INTRODUCTION Recombinant adenoviruses have been the vectors of choice for gene therapy and vaccination due to their high efficiency for gene transfer in a broad spectrum of cell types (Rea et al., 1999; Mountain, 2000). Therefore, market needs for adenovirus vectors (AdV) are constantly increasing causing a high demand of methodologies for the large-scale production of concentrated vectors with warranted levels of purity and efficacy at a low cost (Ferreira et al., 2005a). Different approaches for the improvement of production and purification processes of these viral vectors have mostly used serum-free media (Nadeau et al., 1996; Cote et al., 1997; Cote et al., 1998). Serum-free media formulations are, however, more expensive than the traditional ones with serum supplementation and also the cell specific productivity decreases sharply with increased cell concentration at infection (CCI) above 1×106 cell/ml, the so called “cell density effect” (Nadeau and Kamen 2003; Kamen and Henry 2004). The reasons for such an effect are not very well understood but the lack of an essential nutrient or growth factors or high concentration of metabolic inhibitory products, present at high cell densities, are the most probable causes. Recently, several commercial serum-free media have become available; thus, one of the goals of this study was to evaluate the effect of two different commercially available serum-free media on human 293 cell growth and AdV productivity at high cell densities. Besides culture media composition and the accumulation of inhibitory byproducts, several other factors have been reported to influence cell growth and AdV productivity, namely the vector itself, the cell line, the pH, pCO2 and temperature (Eloit and Adam 1995; Iyer et al., 1999; Jardon and Garnier 2003; Nadeau and Kamen 2003). Due to the common need to implement refeeding strategies for AdV production, it is pertinent to determine if the consequent increase in osmolality has an influence on cell growth and AdV productivity. Only one study has reported the influence of Cell Density Effect 49 osmolality on protein production by human 293 cells (Nadeau et al., 1996), resulting in increased productions at high osmolalities (500 mOsm). Nevertheless, the effect of osmolality on viral productivity has not been assessed yet, thus this study reports effect of osmolality upon human 293 cell growth and AdV productivity. 2. MATERIALS AND METHODS 2.1. Cell line and culture conditions Suspension-adapted human 293 cells (ATCC-CRL-1573) were grown in 125 ml Erlenmeyer flaks, with 40 ml medium in two different serum free media: CD 293, a protein-free medium from Invitrogen and EX-Cell, an animal-protein free medium from JRH, all supplemented with 6 mM glutamine from Invitrogen. Cells were incubated at 37 ºC in a humidified atmosphere of 8 and 10 % of CO2 in air, respectively (in order to obtain an equal pH in each medium), with a cell inoculum of 0.25×106 cell/ml. Regarding the osmolality studies, five different osmolalities were evaluated: 230, 330, 380, 430 and 480 mOsm. The osmolality of the medium was increased by addition of a 1 M NaCl. 2.2. Infection in shake flasks A replication-defective AdV derived from type 5 AdV was kindly provided by Dr. Tom Barret (Institute for Animal Health-Pirbright, UK). Infection was made when the cells reached a concentration of 1×106 cell/ml using a multiplicity of infection (MOI) of 10 (see annex A). Infection was also made when cells reached a concentration of 2×106 cell/ml and 3×106 cell/ml for the studies on media effect, maintaining the same MOI. Infected human 293 cells were harvested at 48 h post infection (hpi), determined as the optimal harvest time (Annex A). Chapter II 50 2.3. Virus culture samples preparation Infected human 293 cells were harvested at 48 hpi and centrifuged at 1000 g for 10 min at 4 ºC; the resulting cell supernatant was distributed into small aliquots and stored at -85 ºC for the evaluation of extracellular virus; the resulting cell pellet was resuspended in a known volume of lysing buffer (Tris/HCl 10 mM, pH 8.0, 2mM MgCl2 and 0.1% Triton-X 100) and the cells were disrupted by 1 minute vortex. Cell debris was removed by centrifugation at 3000 g for 10 min at 4 ºC and the resulting cell lysate supernatant was distributed into small aliquots and stored at –85 ºC for quantification of intracellular viruses content. 2.4. Analytical Methods Cell concentration and viability were determined by counting the cells on a haemacytometer using the Trypan Blue dye exclusion method. AdV titration was performed by the end-point dilution method using 96 well plates and human 293 cells. Infectious particles (ip) in 50 % tissue culture infective dose (TCID50) were determined according to the statistical method of Spearman-Karber (Darling et al., 1998). Ammonia was quantified enzymatically using a UV-test Cat. No. 1112732035 (Boehringer Mannheim, R-Biopharm AG) Osmolality was measured in a Digital Micro Osmometer Type 5R from Hermann Roebling MESSTECHNIK, Germany. 3. RESULTS AND DISCUSSION 3.1. Effect of media on cell growth In order to obtain high cell yields for AdV production, two different serumfree media were tested in their ability to grow human 293 cells. As shown Cell Density Effect 51 in Figure 2.1, EX-Cell allowed a maximum cell number of 7.5×106 cell/ml, while the maximum cell number obtained in CD 293 was approximately 4×106 cell/ml, attained at similar specific growth rates (0.018 h-1 and 0.019 h-1 for CD 293 and EX-Cell, respectively). Despite the fact that EX-Cell has a higher glucose concentration (30 mM) than CD 293 (25 mM), the main difference observed between the two media composition is in the amino acids concentration. By following amino acid consumption it was seen that threonine was the unique amino acid that was completely consumed when cells reached 4×106 cell/ml in CD 293 (data not shown). Furthermore, EXCell contains a concentration of threonine 10 times higher than CD 293, and the refeeding of this amino acid in CD 293 led to an increase in maximum viable cell concentration to 5.5×106 cell/ml (data not shown). Time (days) Viable cells (×10 6 cell/ml) 3 4 5 6 7 8 6 0 1 2 0 1 2 3 4 5 7 8 9 10 11 Time (days) Viable cells (×10 6 cell/ml) 3 4 5 6 7 8 6 0 1 2 0 1 2 3 4 5 7 8 9 10 11 Figure 2.1. Growth curves of human 293 cells obtained in two different culture media: (o) CD 293 and (□) EX-Cell. 3.2. Effect of media on AdV productivity Since EX-Cell medium allowed the higher maximum cell density, the effect upon AdV productivity at high CCI was evaluated and compared with CD 293 medium. As shown in Figure 2.2, when a CCI of 1×106 cell/ml was used, Chapter II 52 both culture media were similar in their ability to “support” AdV production. The specific productivity in human 293 cells using CD 293 medium was not significantly affected when a CCI of 2×106 cell/ml was used, while a 10 fold decrease was observed when EX-Cell medium was used. For a CCI of 3×106 cell/ml no effective AdV production was observed for EX-Cell, and for CD 293 a 1 log decrease (as compared with CCI of 1×106 cell/ml) was observed, confirming the CD 293 medium as a good choice for AdV production at high cell densities (Ferreira et al., 2005b). CCI 1 CCI 2 CCI 3 1E+00 1E+01 1E+02 1E+03 1E+04 CCI 1 CCI 2 CCI 3 ip /Cell CCI 1 CCI 2 CCI 3 1E+00 1E+01 1E+02 1E+03 1E+04 CCI 1 CCI 2 CCI 3 ip /Cell Figure 2.2. Cell specific productivity in CD 293 (■) and EX-Cell (□) at CCIs of 1, 2 and 3×106 cell/ml. Error bars represent the propagated error considering one SD of quadruplicate sample assay for two samples. The decrease in cell specific productivity obtained at a CCI of 3×106 cell/ml in CD 293 and EX-Cell media might be due to the lack of an essential nutrient or growth factors or to the byproducts accumulation. As stated in Chapter III, Part 1, glutamine, threonine and glucose are limiting nutrients in CD 293 medium at CCI of 3×106 cell/ml, although no improvement on AdV productivity was achieved after the addition of these nutrients at the time of infection; on the other hand, ammonia was at an inhibitory concentration for infection at this CCI. In this present study, following ammonia production for both media, it was observed that at 3×106 cell/ml Cell Density Effect 53 ammonia concentration was 1.8 times higher in EX-Cell medium than in CD 293 (Figure 2.3), supporting that ammonia is one of the main contributor for the “cell density effect”. 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Viable cells (10 6 Cell/ml) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0 1 2 3 4 5 6 7 8 Time (h) Ammonia (mM) 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Viable cells (10 6 Cell/ml) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.0 0 1 2 3 4 5 6 7 8 Time (h) Ammonia (mM) Figure 2.3. Ammonia production (full lines) during cell growth (broken lines) in CD 293 (●, ○) and in EX-Cell (■, □). 3.3. Effect of osmolality increase on cell growth and AdV productivity The effect of increasing osmolality on human 293 cell growth was evaluated by increasing it to 330, 380, 430 and 480 mOsm from original values of 230 mOsm at the inoculation time and when the cells reached a concentration of 1×106 cell/ml (concentration where the cell specific productivity is maximal). As shown in Table 2.1, decreases on specific growth rate and maximum viable cell concentration were observed when the osmolality was increased at the innoculation time, leading to growth inhibition at the maximum osmolality studied (480 mOsm), whereas osmolality increases at 1×106 cell/ml had a lesser effect upon both specific growth rate and maximum viable cell concentration. Refeed strategies and non-ammoniagenic medium 61 PART 1 Effect of refeed strategies and non-ammoniagenic medium on adenovirus production at high cell densities Ferreira T.B., Ferreira A.L., Carrondo M.J.T., Alves P.M. (2005) J. Biotechnol., 119:272-280. Chapter III.1 62 Abstract Recombinant adenoviruses became one of the vectors of choice for delivery and expression of foreign proteins for gene therapy and vaccination purposes. Nevertheless, the production of adenovirus is currently limited by the so-called “cell density effect”, i.e., a drop in cell specific productivity concomitant with increased cell concentration at infection (CCI). This work describes the characterisation and optimisation of the infection process in order to improve recombinant adenovirus type 5 yields at high cell densities. For that purpose, 293 cells adapted to suspension were grown in 2 L bioreactors and infected at different cell concentrations, using different refeed strategies, while evaluating cell metabolism. The consumption of amino acids is enhanced during infection, although no amino acid limitation was detected for cells infected at concentrations in the range of 2106 cell/ml, for which the highest volumetric productivity was obtained in batch mode. Conversely, infecting at cell concentrations in the range of 3106 cell/ml led to complete depletion of glucose, glutamine and threonine before the optimal harvesting time, a significant decrease in volumetric productivity being observed; the effect of amino acids and glucose addition at infection time on cell specific and volumetric productivity of adenovirus was assessed, no improvement on adenovirus production being achieved. The effect of ammonia, present in high concentrations at 3106 cell/ml, was evaluated and seem to be detrimental; an 1.8 fold increase on adenovirus volumetric productivity was obtained for infections performed at 3106 cell/ml when nonammoniagenic medium was used. Refeed strategies and non-ammoniagenic medium 63 CONTENTS 1. INTRODUCTION_________________________________________ 64 2. MATERIALS AND METHODS ________________________________ 66 2.1. Cell line and medium ________________________________ 66 2.2. Infection in shake flasks ______________________________ 66 2.3. Stirred tank studies__________________________________ 66 2.4. Virus culture samples preparation ______________________ 67 2.5. Analytical Methods __________________________________ 67 3. RESULTS AND DISCUSSION ________________________________ 68 3.1. Effect of CCI on AdV production ________________________ 68 3.2. Cell metabolism evaluation at different CCIs ______________ 69 3.3. AdV production with a refeed strategy___________________ 72 3.4. AdV production in non-ammoniagenic media ______________ 73 4. CONCLUSIONS__________________________________________ 76 5. ACKNOWLEDGEMENT ____________________________________ 77 6. REFERENCES___________________________________________ 78 Chapter III.1 64 1. INTRODUCTION Delivery of genes to somatic tissue for the prevention and treatment of a wide range of genetic and acquired diseases is a promising route for therapy. The added genes can provide a missing function, modulate the immune response or initiate cell suicide in the presence of certain drugs. Since recombinant adenoviruses (AdV) are highly efficient at gene transfer for a broad spectrum of cell types and species they became one of the vectors of choice for gene delivery and expression of foreign proteins in gene therapy and vaccination (Rea et al., 1999; Mountain 2000). Consequently, the market requirements for AdV are increasing, driving the search for new methodologies for large scale production of concentrated vectors with warranted purity and efficacy. Different approaches for the improvement of production and purification processes of these viral vectors have been reported (Huyghe et al., 1995; Nadeau et al., 1996; Iyer et al., 1999; Blanche et al., 2000). Factors such as the construction of the vector itself, the host cell line and the culture medium can affect AdV production rates (Eloit and Adam 1995; Iyer et al., 1999; Nadeau and Kamen 2003). Effects of pH, pCO2 and temperature on AdV production have also been reported (Xie et al., 2002; Jardon and Garnier 2003). Currently, AdV production at high cell specific productivity is limited to low cell densities, in the range of 1106 cell/ml (Kamen and Henry 2004). This “cell density effect” has been considered by several authors as a consequence of nutrient limitations or accumulation of byproducts. The understanding of these limiting factors will contribute to design a robust and cheap process for AdV production at high cell densities, by adding the essential nutrients or using a new medium formulation that allows the infection at higher cell densities maintaining the specific cell productivity. Although metabolic flux analysis has been used to characterize the effect upon cell metabolism of various environmental conditions the nature of the factors limiting AdV productivities at high cell Refeed strategies and non-ammoniagenic medium 65 densities remains unknown (Nadeau et al., 2000a; Nadeau et al., 2000b; Nadeau et al., 2002). Different strategies for improvement have been proposed: Garnier et al. (1994) demonstrated that medium replacement at 0 hours post infection (hpi) and the addition of glucose at 24 hpi together with periodical pH adjustments, allowed a sustained maximum cell specific productivity at 1.6106 cell/ml whereas, Nadeau et al. (1996) further improved this strategy by also adding essential amino acids at 24 hpi thereby stabilising AdV volumetric productivity at cell densities just above 2106 cell/ml. At production scale, medium exchange will increase the cost of the final product, which is even more critical when the final product is to be used in the veterinary field. Perfusion mode operation has been attempted as a means to control the culture environment and remove toxic byproducts (Henry et al., 2004; Kamen and Henry 2004); nevertheless, the cell specific productivity could only be maintained by infecting cells at densities up to 3106 cell/ml using high perfusions rates of 2 reactor volumes per day, at 2 days post infection, a very costly proposition. The accumulation of ammonia and lactate, major products from glutamine and glucose metabolism, has been demonstrated to diminish mammalian cell growth and influence cell metabolism and protein productivity (Cruz et al., 2000). Attempts have been made to control cell metabolism and reduce the production of these metabolites, such as: i) maintaining low concentrations of glucose and glutamine to shift cell metabolism towards more efficient states (Cruz et al., 1999), or ii) adapting the cells to grow in media with other carbon and nitrogen sources like galactose and glutamate (Altamirano et al., 2000). Specifically, for 293 cells often utilised for AdV production, lactate at 20 mM reduced cell viability to 50% with a much reduced effect on product, whereas ammonia at 1 mM already inhibited cell growth; under these conditions the effects upon AdV production were not identified (Nadeau et al., 1996). In this work, factors limiting the AdV production at high cell densities were Chapter III.1 66 investigated. Refeed strategies and the adaptation to non-ammoniagenic media were evaluated in order to understand the metabolic requirements when infection is done at high cell densities and to overcome the so called "cell density effect", thus permitting the achievement of higher volumetric productivities for AdV. A first generation AdV vector was employed. 2. MATERIALS AND METHODS 2.1. Cell line and medium 293 cells, purchased from ATCC (ATCC-CRL-1573), were adapted to suspension and grown in commercially available serum and protein free medium, CD 293, supplemented with 4 mM of glutamine (all from Invitrogen, Glasgow, UK) at a humidified atmosphere of 8% CO2 in air at 37ºC in shake flasks (corning, NY). Cells were routinely propagated twice a week with an inoculum of 0.5×106 cell/ml. 2.2. Infection in shake flasks Infection was performed in 125 ml shake flasks with final volumes of 40 ml in a humidified atmosphere of 8% CO2 in air at 37ºC. All the infections were done using a multiplicity of infection (MOI) of 10 and AdV harvested at 48 hpi (see Annex A). 2.3. Stirred tank studies Bioreaction studies were performed in a 2 L bioreactor (Braun, Melsungen, Germany). The agitation rate was maintained at 110 rpm (NRe = 6800); pH was controlled at 7.2 by aeration with a CO2 gas-mixture and NaOH 0.2 M; the dissolved oxygen was controlled at 80% air saturation. All the infections were done using a MOI of 10 and AdV harvested at 48 hpi (see Annex A). For Refeed strategies and non-ammoniagenic medium 67 the fed-batch operation mode, threonine, glutamate, amino acids cocktail (RPMI 1640 50) (all from Sigma-Aldrich, St. Louis, MO), glutamine and glucose (Merck, Darmstadt, Germany) solutions were used. 2.4. Virus culture samples preparation A replication-defective AdV derived from type 5 AdV was kindly provided by Dr. Tom Barret (IAH-Pirbright, UK). Infected 293 cells were harvested at 48 hpi and centrifuged at 1000g for 10 min. at 4ºC. The resulting cell supernatant was distributed into small aliquots and stored at -85ºC for the evaluation of extracellular virus. The resulting cell pellet was resuspended in a known volume of Tris buffer 10 mM, pH 8.0 supplemented with 2 mM MgCl2 and 0.1% Triton. Cells were disrupted by vortexing for 1 minute and cell debris removed by centrifugation at 3000g for 10 min. at 4ºC. The resulting cell lysate supernatant was dispensed into small aliquots and stored at -85ºC for quantification of intracellular viral content. 2.5. Analytical Methods Cell concentration and viability were determined by counting cells on a Fuchs-Rosenthal haemocytometer (Brand, Wertheim, Germany) using the trypan blue (Invitrogen) dye exclusion method. AV titration was performed by the end-point dilution method (TCID50) using 96 well plates and 293 cells. The titer was determined according to the method of Spearman and Kraber, as described elsewhere (Darling et al., 1998). Glucose and lactate were analyzed using a YSI Multiparameter Bioanalytical System Model 7100 MBS (Yellow Springs, USA). Ammonia was quantified enzymatically using a UV test number 1112732035 (Boehringer Mannheim, R-Biopharm AG, Germany). Chapter III.1 68 Complete analysis of amino acid composition was done by HPLC using a Pico-Tag (Waters) system. Amino acids were derivatized with the phenylisothiocianate method (Matsudaira 1990). The Pico-Tag system consisted of a Nova-Pak C18 column (WAT 011695), a Waters 510 pump, and automatic injector Waters 717 and a tunable Absorbance Detector 486. Data acquisition was performed on a PC using software from Millenium. 3. RESULTS AND DISCUSSION 3.1. Effect of CCI on AdV production The effect of CCI on the final AdV production yields was evaluated in 2 L bioreactor. Infection at CCI 1106 cell/ml (control), 2106 cell/ml and 3106 cell/ml with or without complete media exchange at the time of infection was compared (Fig. 3.1.1). As can be observed, cell specific productivity is not significantly affected at CCI 2106 cell/ml, when compared with the control, leading to a 1.3 fold increase on AdV volumetric productivity. However, one log decrease on cell specific productivity was observed at CCI 3106 cell/ml. Medium exchange at the time of infection and the addition of glucose and essential amino acids at 24 hpi has been reported to allow a sustained maximum specific productivity just above 2106 cell/ml (Nadeau et al., 1996). Figure 3.1.1 shows that a single medium exchange at the time of infection is enough to maintain the cell specific productivity at the maximum CCI of 3106 cell/ml, for which the highest volumetric productivity was obtained. The results obtained are improvements on those reported in the literature, suggesting the serum and protein free medium used in this study (CD 293) to be a good choice for AdV production at high cell densities. Nevertheless, at production scale, a medium exchange could make the final product cost prohibitive. To determine how these results Refeed strategies and non-ammoniagenic medium 69 are related to the metabolism of the cells, the kinetics of nutrient consumption and metabolite production were evaluated. 1E+01 1E+02 1E+03 1E+04 CCI 1 CCI 1+ media exchange CCI 2 CCI 2+ media exchange CCI 3 CCI 3 + media exchange ip/cell 1E+08 1E+09 1E+10 1E+11 ip/ml 1E+01 1E+02 1E+03 1E+04 CCI 1 CCI 1+ media exchange CCI 2 CCI 2+ media exchange CCI 3 CCI 3 + media exchange ip/cell 1E+08 1E+09 1E+10 1E+11 ip/ml Figure 3.1.1. Effect of media exchange and CCI in cell specific ( ) and AdV volumetric ( ) productivities. CCI 1, CCI 2 and CCI 3 correspond to 1, 2 and 3106 cell/ml at time of infection, respectively. 3.2. Cell metabolism evaluation at different CCIs Cell metabolism was assessed before and after infection at CCI 1106 cell/ml and 3106 cell/ml. After infection at CCI 1106 cell/ml, increases of 1.3 fold in glutamine and 1.8 fold in glucose specific consumption rates were observed, concomitant with increases of 1.2 fold in ammonia and 3.9 fold in lactate specific production rates (data not shown). No significant alteration in the molar ratio of ammonia produced/glutamine consumed was observed, whereas the molar ratio of lactate produced/glucose consumed increased 2.3 times (Table 3.1.1). Elias et al. (2003) observed that for 293 cells, the flux of glucose into the TCA cycle and its subsequent utilization is limited as a result of the lack of certain key enzymes in this pathway. This may be the explanation for the significant increase in the molar ratio of lactate produced/glucose consumed observed in our Chapter III.1 76 ip/cell ip/ml 1E+08 1E+09 1E+10 1E+11 1E+01 1E+02 1E+03 1E+04 Gln CCI 1 Gln CCI 3 Glu CCI 1 Glu CCI 3 ip/cell ip/ml 1E+08 1E+09 1E+10 1E+11 1E+01 1E+02 1E+03 1E+04 Gln CCI 1 Gln CCI 3 Glu CCI 1 Glu CCI 3 Figure 3.1.7. Effect of medium supplement in cell specific ( ) and AdV volumetric () productivities. Gln and Glu correspond to glutamine and glutamate supplemented medium and CCI 1 and CCI 3 correspond to 1 and 3106 cell/ml at time of infection, respectively. These results suggest an inhibitory effect on AdV production by ammonia accumulation and confirmed that the use of non-ammoniagenic medium improved AdV volumetric productivity. However, the ammonia accumulation seems not to be the only inhibitory factor on AdV production, since in optimal condition, a 3 times increase on AdV volumetric productivity was to be expected at CCI 3×106 cell/ml, when only an 1.8 fold increase was obtained, corresponding to the highest decrease in cell specific productivity indicated above. 4. CONCLUSIONS The present study shows that by using CD 293 medium the AdV production reaches the highest volumetric productivity reported for an infection performed at CCI of 2106 cell/ml in batch mode. Moreover, AdV productivity can be improved even at a CCI of 3106 cell/ml with a single Refeed strategies and non-ammoniagenic medium 77 CD 293 medium exchange at the time of infection. Nevertheless, for scaled up processes, a medium exchange step is not cost effective. Glutamine, threonine and glucose were found to be limiting nutrients at CCI of 3106 cell/ml, although no improvement on AdV productivity could be achieved after the addition of these nutrients at the time of infection. On the other hand, ammonia was at an inhibitory concentration for infection at this CCI. By adapting the cells to non-ammoniagenic medium, an 1.8 fold increase in AdV volumetric productivity at CCI 3106 cell/ml was obtained, although the “ideal” situation would have been an increase of 3 times at this CCI; thus, ammonia is an important parameter to be considered for infection at high cell densities, but not the only. Future improvements to the culture process may come from analysis of other nutrients than the usual glucose, lactate, ammonia and amino acids. A cocktail of different nutrients are present in the culture media that may have an important impact on AdV production as vitamins, lipids, hormones and growth factors (see Annex B). Combination of these measurements may lead to the design of a new strategy permitting a higher cell specific and, consequently, AdV volumetric productivity. 5. ACKNOWLEDGEMENT The authors are grateful to Dr Tom Barret (IAH-Pirbright, UK) for providing the recombinant adenovirus and Eng Paula Chicau for providing data from the amino acids analysis service at the Instituto de Tecnologia Química e Biológica, Universidade Nova de Lisboa, Oeiras, Portugal. The authors acknowledge and appreciate the financial support received from the European Commission (Project ICF599A4PR01) and from Fundação para a Ciência e Tecnologia – Portugal (Project POCTI/BIO/46515/2002 and student grant SFRH/BD/10614/2002). Chapter III.1 78 6. REFERENCES Altamirano, C., Paredes, C., Cairo, J.J. and Godia, F., 2000. Improvement of CHO cell culture medium formulation: simultaneous substitution of glucose and glutamine. Biotechnol Prog 16, 69-75. Blanche, F., Cameron, B., Barbot, A., Ferrero, L., Guillemin, T., Guyot, S., Somarriba, S. and Bisch, D., 2000. An improved anion-exchange HPLC method for the detection and purification of adenoviral particles. Gene Ther 7, 1055-1062. Christie, A. and Butler, M., 1999. The adaptation of BHK cells to a nonammoniagenic glutamate-based culture medium. Biotechnol Bioeng 64, 298-309. Cruz, H.J., Ferreira, A.S., Freitas, C.M., Moreira, J.L. and Carrondo, M.J., 1999. Metabolic responses to different glucose and glutamine levels in baby hamster kidney cell culture. Appl Microbiol Biotechnol 51, 579-585. Cruz, H.J., Freitas, C.M., Alves, P.M., Moreira, J.L. and Carrondo, M.J., 2000. Effects of ammonia and lactate on growth, metabolism, and productivity of BHK cells. Enzyme Microb Technol 27, 43-52. Darling, A.J., Boose, J.A. and Spaltro, J., 1998. Virus assay methods: accuracy and validation. Biologicals 26, 105-10. Elias, C.B., Carpentier, E., Durocher, Y., Bisson, L., Wagner, R. and Kamen, A., 2003. Improving glucose and glutamine metabolism of human HEK 293 and Trichoplusia ni insect cells engineered to express a cytosolic pyruvate carboxylase enzyme. Biotechnol Prog 19, 90-97. Eloit, M. and Adam, M., 1995. Isogenic adenoviruses type 5 expressing or not expressing the E1A gene: efficiency as virus vectors in the vaccination of permissive and non-permissive species. J Gen Virol 76, 1583-1589. Garnier, A., Cote, J., Nadeau, I., Kamen, A. and Massie, B., 1994. Scale-up of the adenovirus expression system for the production of recombinant protein in human 293S cells. Cytotechnology 15, 145-155. Refeed strategies and non-ammoniagenic medium 79 Henry, O., Dormond, E., Perrier, M. and Kamen, A., 2004. Insights into adenoviral vector production kinetics in acoustic filter-based perfusion cultures. Biotechnol Bioeng 86, 765-774. Huyghe, B.G., Liu, X., Sutjipto, S., Sugarman, B.J., Horn, M.T., Shepard, H.M., Scandella, C.J. and Shabram, P., 1995. Purification of a type 5 recombinant adenovirus encoding human p53 by column chromatography. Hum Gene Ther 6, 1403-16. Iyer, P., Ostrove, J.M. and Vacante, D., 1999. Comparison of manufacturing techniques for adenovirus production. Cytotechnology 30, 169-172. Jardon, M. and Garnier, A., 2003. pH, pCO2, and Temperature Effect on RAdenovirus Production. Biotechnol Prog 19, 202-208. Kamen, A. and Henry, O., 2004. Development and optimization of an adenovirus production process. J Gene Med 6 Suppl 1, S184-192. Matsudaira, P. 1990. Methods in enzymology. Press, A., editor. San Diego. McQueen, A. and Bailey, J.E., 1991. Growth inhibition of hybridoma cells by ammonium ion: correlation with effects on intracellular pH. Biopr Eng 6, 49-61. Mountain, A., 2000. Gene therapy: the first decade. Trends Biotechnol 18, 119-128. Nadeau, I., Garnier, A., Côté, J., Massie, B., Chavarie, C. and Kamen, A., 1996. Improvement of recombinant protein production with human adenovirus/293S expression system using fed-batch Strategies. Biotechnol Bioeng 51, 613-623. Nadeau, I., Gilbert, P.A., Jacob, D., Perrier, M. and Kamen, A., 2002. Low-protein medium affects the 293SF central metabolism during growth and infection with adenovirus. Biotechnol Bioeng 77, 91-104. Nadeau, I., Jacob, D., Perrier, M. and Kamen, A., 2000a. 293SF Metabolic Flux analysis during cell growth and infection with an adenoviral vector. Biotechnol Prog 16, 872-884. Chapter III.1 80 Nadeau, I. and Kamen, A., 2003. Production of adenovirus vector for gene therapy. Biotechnol Adv 20, 475-489. Nadeau, I., Sabatie, J., Koehl, M., Perrier, M. and Kamen, A., 2000b. Human 293 cell metabolism in low glutamine-supplied culture: interpretation of metabolic changes through metabolic flux analysis. Metab Eng 2, 277-292. Rea, D., Schagen, F.H., Hoeben, R.C., Mehtali, M., Havenga, M.J., Toes, R.E., Melief, C.J. and Offringa, R., 1999. Adenoviruses activate human dendritic cells without polarization toward a T-helper type 1-inducing subset. J Virol 73, 1024510253. Xie, L., Pilbrough, W., Metallo, C., Zhong, T., Pikus, L., Leung, J., Aunins, J.G. and Zhou, W., 2002. Serum-free suspension cultivation of PER.C6R cells and recombinant adenovirus production under different pH conditions. Biotechnol Bioeng 80, 569-579. Ammonia effect on pHi and AdV production 81 PART 2 Effect of ammonia production on intracellular pH: consequent effect on adenovirus vector production Ferreira T.B., Carrondo M.J.T., Alves P.M. (2007) J. Biotechnol., 119:272280. Chapter III.2 82 Abstract Recombinant adenoviral vectors (AdV) have proven to be highly efficient for the delivery and expression of foreign genes in a broad spectrum of cell types and species both for vaccination and gene therapy in a number of specific applications. In this study, the effect of ammonia production on intracellular pH (pHi) and consequently inhibition of AdV production at high cell densities is assessed. Different specific ammonia production rates were obtained for 293 cells adapted to grow in glutamate supplemented medium (nonammoniagenic medium) as compared with 293 cells growing in glutamine supplemented medium (ammoniagenic medium); pHi was observed to be lower during cell growth and AdV production at both high and low CCI in the ammoniagenic medium, where the specific ammonia production rate is higher. In addition, after infection at CCI of 3106 cell/ml, the cell viability decreased significantly in the ammoniagenic medium, attributed to the activation of an acidic pathway of apoptosis. Furthermore, AdV DNA was observed to be degraded at the observed pHi in the ammoniagenic medium, decreasing significantly the amount of AdV DNA available for encapsidation. To elucidate the pHi effect upon AdV production, 293 cells were infected at a CCI of 1106 cell/ml in the non-ammoniagenic medium with a manipulated pHi as observed at the time of infection at CCI of 3106 cell/ml in the ammoniagenic (pHi 7.0) and non-ammoniagenic (pHi 7.3) media; AdV volumetric productivities were observed to be lower when the cells were exposed to the lower pHi. Thus, the importance of controlling all the factors contributing to pHi on AdV production, such as ammonia production, has been established. Ammonia effect on pHi and AdV production 83 CONTENTS 1. INTRODUCTION_________________________________________ 84 2. MATERIALS AND METHODS ________________________________ 86 2.1. Cell line and medium ________________________________ 86 2.2. Infection in shake flasks ______________________________ 87 2.3. Virus culture samples preparation ______________________ 87 2.4. Adenoviral vector titration ____________________________ 88 2.5. Intracellular pH measurement _________________________ 88 2.6. DNase II activity assay at different pH values______________ 89 2.7. Determination of adenovirus DNA by quantitative real time PCR _____________________________________________________ 89 2.8. Analytical Methods __________________________________ 90 3. RESULTS AND DISCUSSION ________________________________ 90 3.1. Effect of ammonia addition on adenovirus vector production _ 90 3.2. Effect of ammonia on intracellular pH during cell growth ____ 91 3.3. Evaluation of ammonia production and intracellular pH during adenovirus vector production at different CCIs ________________ 93 3.4. DNase II effect upon adenovirus vector DNA at different pHs__ 96 3.5. Effect of intracellular pH on adenovirus vector production ___ 97 4. CONCLUSIONS__________________________________________ 98 5. ACKNOWLEDGEMENT ____________________________________ 98 6. REFERENCES__________________________________________ 100 Chapter III.2 84 1. INTRODUCTION Recombinant adenoviral vectors (AdV) have proven to be highly efficient for the delivery and expression of foreign genes in a broad spectrum of cell types and species, in a number of specific applications (Volpers and Kochanek, 2004; Ferreira et al., 2005a). Thus, as the market needs for AdV are increasing, methodologies for production of concentrated AdV are needed. Although 293 cells can be grown up to 8106 cell/ml in batch systems (Ferreira et al., 2005c), AdV production at high cell densities is limited by the so-called “cell density effect”, i.e., a drop in cell specific productivity concomitant with increased cell concentration at infection (CCI) for values above 1106 cells/ml (Nadeau and Kamen, 2003; Kamen and Henry, 2004; Ferreira et al., 2005b). Although this effect has been attributed mainly to nutrient limitations and/or byproducts accumulation, the exact nature of the factors limiting AdV productivities at high CCIs remains unknown. Previous studies have indicated that medium replacement at the time of infection and the addition of glucose and amino acids at 24 h post infection (hpi) together with periodic pH adjustments, allows for high cell specific productivity at cell densities in the range 2106 to 3106 cells/ml (Garnier et al., 1994; Nadeau et al., 1996). Our group has shown that ammonia is an important parameter for infection at high cell densities, since, by adapting 293 cells to non-ammoniagenic medium, a 5 fold increase on cell specific productivity could be obtained at a CCI of 3×106 cell/ml (Ferreira et al., 2005b). However, the mechanisms by which ammonia inhibits AdV production remain unknown. Ammonia derives mainly from deamination of glutamine, whose concentration in the culture medium is usually 5 to 20 fold higher than that of other amino acids (Eagle, 1955). Complete catabolism of glutamine produces 2 mol of ammonia per mol of glutamine. Other minor sources of ammonia include the spontaneous degradation of glutamine (Ozturk and Palsson, 1990; Arii et al., 1999) and catabolism of other amino acids. Ammonia effect on pHi and AdV production 85 Several mechanisms of ammonia toxicity have been described in animal cells, including: disturbance of electrochemical gradients (Mirabet et al., 1997); inhibition of enzyme reactions; changes in intracellular pH (pHi) (McQueen and Bailey, 1990); increased demand for energy maintenance (Martinelle and Haggstrom, 1993); inhibition of cell growth; and perturbed processing and secretion of proteins and apoptosis induction (Cruz et al., 2000). H+ - pH H+ H+ +NH4+H+NH3 + NH4+H+NH3 + - Na+ Na+ osmolality +Na+ +NH4+H+NH3 + NH4+H+NH3 + Cytoplasm Culture Medium Cell membrane H+ - pH H+ H+ H+ H+ +NH4+H+NH3 + NH4+H+NH3 + NH4+H+NH3 + NH4+H+NH3 + - Na+ Na+ osmolality +Na+ +NH4+H+NH3 + NH4+H+NH3 + NH4+H+NH3 + NH4+H+NH3 + Cytoplasm Culture Medium Cell membrane Figure 3.2.1 Mechanisms influencing pHi at decreased medium pH, increased ammonium ion concentration, and increased medium osmolality (adapted from Cherlet and Marc (1998)). In the case of pHi, ammonia can disturb the optimal cytoplasmatic pH by forming a “proton shuttle” (represented schematically in Figure 3.2.1): briefly, the intracellularly originated ammonium ions (NH4+) are rapidly excreted by passive diffusion, under the form of ammonia ions (NH3), to the extracellular milieu with concomitant accumulation inside the cell of one proton (H+) per NH3 molecule released. This excretion is facilitated by the negative charge of the membrane potential and the gradient of ammonia ions across the cell membrane. Once outside the cell, NH4+can enter the cell again propelled by the large electrical gradient due to the highly charged membrane and leave the cell again under the NH3 form; this cycle can be started over and over again (McQueen and Bailey, 1990; Wu et Chapter III.2 92 medium (qammonia = 410-9 mmol/cell.h) in terms of growth, ammonia production and pHi. As shown in Figure 3.2.3, for the non-ammoniagenic medium, the cells were able to maintain the pHi between 7.2 and 7.3 throughout all the culture; for ammoniagenic medium the pHi was considerably lower: a drop of 0.2 units was evident at time 0 h and, as cells reached the concentration of 2106 cell/ml (at 96 h), a significant decrease in pHi was registered. For the non-ammoniagenic medium, the pHe was maintained at 7.20 during all the culture time. For the ammoniagenic medium, a decrease of 0.05 (from 7.20 to 7.15) was observed from time 0 to time 96; however, this decrease was not as significant as the one observed for the pHi. 0E+00 1E+06 2E+06 3E+06 4E+06 5E+06 6E+06 0 24 48 72 96 120 144 168 192 216 240 time (h) Viable cells (cell/ml) 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi 0E+00 1E+06 2E+06 3E+06 4E+06 5E+06 6E+06 0 24 48 72 96 120 144 168 192 216 240 time (h) Viable cells (cell/ml) 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi Figure 3.2.3. Evaluation of pHi (empty symbols) during 293 cells growth (full symbols) in ammoniagenic medium (circles) and non-ammoniagenic medium (triangles). By monitoring ammonia concentration it is possible to correlate the changes observed in pHi with the ammonia production (Figure 3.2.4). As expected, for the non-amoniagenic medium the production of ammonia Ammonia effect on pHi and AdV production 93 was significantly lower than for the ammoniagenic medium. The observed delay in pHi decrease relative to the faster enhancement of ammonia concentration in the culture medium can be related with the increasing number of “proton shuttle” cycles (McQueen and Bailey, 1990; Wu et al., 1993) leading to a much higher uptake of protons then the overall cell capacity to excrete them. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 24 48 72 96 120 144 168 192 216 240 time (h) Ammonia (mM) 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 24 48 72 96 120 144 168 192 216 240 time (h) Ammonia (mM) 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi Figure 3.2.4. Relationship between pHi (empty symbols) and ammonia production (full symbols) in ammoniagenic medium (circles) and non-ammoniagenic medium (triangles). 3.3. Evaluation of ammonia production and intracellular pH during adenovirus vector production at different CCIs In order to evaluate the effect of produced ammonia on pHi during the production of AdV at a CCI of 1 and 3×106 cell/ml, 293 cells were infected in either non-ammoniagenic and ammoniagenic media. As can be observed in Figure 3.2.5, the pHi is severely affected after infection, independently of the medium and CCI used. However, for both culture media tested, this Chapter III.2 94 decrease is more pronounced when the infection is performed at CCI of 3×106 cell/ml, especially for ammoniagenic media. 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 6 12 18 24 30 36 42 48 time (hpi) Ammonia (mM) 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi CCI = 1×106cell/ml 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 6 12 18 24 30 36 42 48 time (hpi) Ammonia (mM) 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi CCI = 1×106cell/ml 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 6 12 18 24 30 36 42 48 time (hpi) Ammonia (mM) 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi CCI = 3×106cell/ml 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 0 6 12 18 24 30 36 42 48 time (hpi) Ammonia (mM) 5.7 5.8 5.9 6.0 6.1 6.2 6.3 6.4 6.5 6.6 6.7 6.8 6.9 7.0 7.1 7.2 7.3 pHi CCI = 3×106cell/ml Figure 3.2.5. Evaluation of pHi (empty symbols) and ammonia (full symbols) during AdV productions at CCI of 1106 cell/ml in ammoniagenic medium (squares) and non-ammoniagenic medium (diamonds) and CCI of 3106 cell/ml in ammoniagenic medium (circles) and non-ammoniagenic medium (triangles) with a MOI of 10. Ammonia effect on pHi and AdV production 95 The specific ammonia production rate was not significantly affected when the CCI increased from 1 to 3×106 cell/ml, thus the observed pHi decrease (Figure 3.2.5) seems to be promoted by other factors than the ammonia production. Matsuyama et al. (2000) shown that a very early event in apoptosis involves a change on the pHi regulation, meaning that the significant decrease observed in pHi after AdV infection may be a way to the cell respond to infection by entering in apoptosis. In addition, Matsuyama et al. (2000) have shown that the efficiency of caspase activation, responsible for the beginning of the apoptosis process, is pH sensitive, with an in vitro pH optimum of approximately 6.6. In fact, the amount of viable cells after infection at CCI of 3106 cell/ml start to undergo a more pronounced decrease when the pHi reaches the value of 6.6 (Figure 3.2.5) at 24 hpi for the ammoniagenic medium versus 36 hpi for the non-ammoniagenic medium (Figure 3.2.6); this means that cells are kept viable after infection for a longer time when the non-ammoniagenic medium is used. 0.0E+00 5.0E+05 1.0E+06 1.5E+06 2.0E+06 2.5E+06 3.0E+06 3.5E+06 0 6 12 18 24 30 36 42 48 time (hpi) Viable cells (cell/ml) 0 5 10 15 20 25 30 % dead cells 0.0E+00 5.0E+05 1.0E+06 1.5E+06 2.0E+06 2.5E+06 3.0E+06 3.5E+06 0 6 12 18 24 30 36 42 48 time (hpi) Viable cells (cell/ml) 0 5 10 15 20 25 30 % dead cells Figure 3.2.6. Effect of CCI of 3106 cell/ml in cell viability (full symbols) and percentage of dead cells (empty symbols) in ammoniagenic medium (circles) and Chapter III.2 96 non-ammoniagenic medium (triangles). 3.4. DNase II effect upon adenovirus vector DNA at different pHs To evaluate how pHi affects DNase II activity and consequently AdV DNA, incubations of the viral DNA were done at different pH values. As shown in Figure 3.2.7, pH affects the activity of DNase II upon the AdV DNA, a significant decline being observed for pH values below 6.6. No significant effects in AdV DNA were observed for similar studies performed with DNase I (data not shown). 1E+03 1E+04 1E+05 1E+06 1E+07 I0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 pH DNA (copies/ml) 1E+03 1E+04 1E+05 1E+06 1E+07 I0 7.8 7.6 7.4 7.2 7.0 6.8 6.6 6.4 6.2 6.0 5.8 5.6 pH DNA (copies/ml) Figure 3.2.7. Effect of pH on the activity of DNase II. AdV DNA was incubated with a cell supernatant extract as described in material and methods. () Represents the initial amount of AdV DNA observed at time 0 for each pH tested. Comparing the activity of DNase II with the pHi profiles after infection at a CCI of 1 and 3106 cell/ml in the ammoniagenic and non-ammoniagenic medium (Figure 3.2.5), it is possible to conclude that at CCI of 1106 cell/ml, independently of the medium used, only at the end of the production process the pHi reach values below 6.6. On the other hand, when cells are infected at CCI of 3106 cell/ml in the ammoniagenic medium, the pHi reach the 6.6 units faster than in the non-ammoniagenic Ammonia effect on pHi and AdV production 97 medium. Thus, for ammoniagenic media, due to the significant increase on DNase II activity for values below pH 6.6, more AdV DNA degradation will occur leading to less availability of virus DNA for encapsidation with a consequent decrease in bioactive AdV production. 3.5. Effect of intracellular pH on adenovirus vector production To elucidate the pHi effect upon AdV production, 293 cells adapted to grow in the non-ammoniagenic medium were infected at a CCI of 1106 cell/ml at pH 7.0 or at pH 7.3 (pHi observed at the time of infection at CCI of 3106 cell/ml in the ammoniagenic and non-ammoniagenic media respectively), with or without nigericin addition and harvested at 48 hpi. Nigericin was used to mimic the pHi observed at CCI 3106 cell/ml in both media at pHi 7.0 and 7.3. 1E+07 1E+08 1E+09 1E+10 pH 7.0 (control) pH 7.3 (control) pH 7.0 pH 7.3 ip/ml 1E+07 1E+08 1E+09 1E+10 (control) (control) ip/ml 1E+07 1E+08 1E+09 1E+10 pH 7.0 (control) pH 7.3 (control) pH 7.0 pH 7.3 ip/ml 1E+07 1E+08 1E+09 1E+10 (control) (control) ip/ml Figure 3.2.8. Effect of pHi on AdV volumetric productivity. pHi was decreased by addition of 30 mM KCl and of 50 M nigericin as described in material and methods. As can be observed in Figure 3.2.8, although there is no significant difference on AdV production at pH 7.0 or at pH 7.3 in control conditions, Chapter III.2 98 when nigericin is used a drop in AdV volumetric productivity is observed for both pHs, although it is more pronounced at pH 7.0. Despite the fact that nigericin has a residual cytotoxic effect (data not shown), it is obvious that, for a pH of 7.0, AdV volumetric productivity is strongly affected. These results reflect the impact of the pHi on AdV production and the overall need to minimize all the factors that contribute to its decrease, such as the intracellular ammonia production. 4. CONCLUSIONS The present study compares the effect of two different culture media leading to different specific ammonia production rates, upon pHi and AdV productivity in 293 cells infected at different CCIs (1 and 3×106 cell/ml). During cell growth the pHi decreased significantly in the ammoniagenic medium, where the specific ammonia production rate is higher; furthermore, after infection, the pHi dropped significantly at both CCIs for both media, although the lowest pHi was obtained for the ammoniagenic medium. These changes observed in the pHi were found to play an important role on both processes of apoptosis and AdV DNA degradation, leading to a decrease in bioactive AdV production, more severe for the ammoniagenic medium. The data reported herein identifies pHi as one of the key factors responsible for the “cell density effect” observed on AdV production. Nonammoniagenic medium was shown as a good option to be implemented at the bioreaction level that will lead to the maintenance of cell viability during AdV production and consequent increase on cell specific productivity. 5. ACKNOWLEDGEMENT The authors are grateful to Dr Tom Barret (Institute for Animal Health- Ammonia effect on pHi and AdV production 99 Pirbright, UK) for providing the recombinant adenovirus, Eng Marlene Carmo for all the support concerning the flow cytometric analyzes and Eng Rosário Clemente for PCR analysis. The authors acknowledge and appreciate the financial support received from the European Commission (Project RP/PPR ORALVAC ICA4-CT-2000-30027) and from Fundação para a Ciência e Tecnologia, Portugal (Project POCTI/BIO/46515/2002 and student grant SFRH/BD/10614/2002). Chapter III.2 100 6. REFERENCES Arii, K., Kobayashi, H., Kai, T., and Kokuba, Y., 1999. Degradation kinetics of Lglutamine in aqueous solution. Eur J Pharm Sci 9, 75-8. Barry, M. A., and Eastman, A., 1993. Identification of deoxyribonuclease II as an endonuclease involved in apoptosis. Arch Biochem Biophys 300, 440-50. Cherlet, M., and Marc, A., 1998. Intracellular pH monitoring as a tool for the study of hybridoma cell behavior in batch and continuous bioreactor cultures. Biotechnol Prog 14, 626-38. Cruz, H. J., Freitas, C. M., Alves, P. M., Moreira, J. L., and Carrondo, M. J., 2000. Effects of ammonia and lactate on growth, metabolism, and productivity of BHK cells. Enzyme Microb Technol 27, 43-52. Eagle, H., 1955. Nutrition needs of mammalian cells in tissue culture. Science 122, 501-14. Evans, R. K., Nawrocki, D. K., Isopi, L. A., Williams, D. M., Casimiro, D. R., Chin, S., Chen, M., Zhu, D. M., Shiver, J. W., and Volkin, D. B., 2004. Development of stable liquid formulations for adenovirus-based vaccines. J Pharm Sci 93, 2458-75. Ferreira, T. B., Alves, P. M., Aunins, J. G., and Carrondo, M. J. T., 2005a. Use of adenoviral vectors as veterinary vaccines. Gene Ther 12 Suppl 1, S73-83. Ferreira, T. B., Ferreira, A. L., Carrondo, M. J. T., and Alves, P. M., 2005b. Effect of refeed strategies and non-ammoniagenic medium on adenovirus production at high cell densities. J Biotechnol 119, 272-80. Ferreira, T. B., Ferreira, A. L., Carrondo, M. J. T., and Alves, P. M., 2005c. Two different serum-free media and osmolality effect upon 293 cell growth and adenovirus production. Biotechnol Lett 27, 1809-13. Frelin, C., Vigne, P., Ladoux, A., and Lazdunski, M., 1988. The regulation of the intracellular pH in cells from vertebrates. Eur J Biochem 174, 3-14. Ammonia effect on pHi and AdV production 101 Garnier, A., Cote, J., Nadeau, I., Kamen, A., and Massie, B., 1994. Scale-up of the adenovirus expression system for the production of recombinant protein in human 293S cells. Cytotechnology 15, 145-155. Gottlieb, R. A., Nordberg, J., Skowronski, E., and Babior, B. M., 1996. Apoptosis induced in Jurkat cells by several agents is preceded by intracellular acidification. Proc Natl Acad Sci U S A 93, 654-8. Kamen, A., and Henry, O., 2004. Development and optimization of an adenovirus production process. J Gene Med 6 Suppl 1, S184-192. Lagadic-Gossmann, D., Huc, L., and Lecureur, V., 2004. Alterations of intracellular pH homeostasis in apoptosis: origins and roles. Cell Death Differ 11, 953-61. Martinelle, K., and Haggstrom, L., 1993. Mechanisms of ammonia and ammonium ion toxicity in animal cells: transport across cell membranes. J Biotechnol 30, 33950. Matsuyama, S., Llopis, J., Deveraux, Q. L., Tsien, R. Y., and Reed, J. C., 2000. Changes in intramitochondrial and cytosolic pH: early events that modulate caspase activation during apoptosis. Nat Cell Biol 2, 318-25. McQueen, A., and Bailey, J. E., 1990. Effect of Ammonium Ion and Extracellular Ph on Hybridoma Cell-Metabolism and Antibody-Production. Biotechnol Bioeng 35, 1067-1077. Mirabet, M., Navarro, A., Lopez, A., Canela, E. I., Mallol, J., Lluis, C., and Franco, R., 1997. Ammonium toxicity in different cell lines. Biotechnology and Bioengineering 56, 530-537. Nadeau, I., Garnier, A., Côté, J., Massie, B., Chavarie, C., and Kamen, A., 1996. Improvement of recombinant protein production with human adenovirus/293S expression system using fed-batch Strategies. Biotechnol Bioeng 51, 613-623. Nadeau, I., and Kamen, A., 2003. Production of adenovirus vector for gene therapy. Biotechnol Adv 20, 475-489. Chapter IV.1 108 1. INTRODUCTION AdV have been extensively used as vectors for both gene therapy and recombinant DNA vaccines, which have been tested in both humans and different animal species (Ferreira et al., 2005a). The need for large quantities of clinical-grade AdV is an important limitation to in vitro experimentations as well as to pre-clinical and clinical studies (Nadeau and Kamen 2003; Kamen and Henry 2004). This limitation is due to the fact that the cell specific productivity decreases sharply with increased CCI above 1106 cell/ml, the so called “cell density effect”. So far, this “cell density effect” has been attributed to nutrient limitations and/or accumulation of byproducts (Nadeau and Kamen 2003; Henry et al., 2004; Kamen and Henry 2004; Ferreira et al., 2005b). Although metabolic flux analysis has been used to characterize the effect of various environmental conditions upon cell metabolism, the exact nature of the factors limiting AdV productivities at high cell density remains unknown (Nadeau et al., 2000a; Nadeau et al., 2000b; Nadeau et al., 2002). Therefore, present strategies for overcoming this bottleneck have been focused on the improvement of cell growth environment during and after infection. These strategies include medium replacement at the time of infection and the addition of glucose at 24 hour post infection (hpi) together with periodical pH adjustments, allowing a sustained maximum specific productivity at 1.6106 cell/ml (Garnier et al., 1994). This strategy, associated with essential amino acids feeding at 24 hpi, is able to stabilise the volumetric productivity at cell densities up to 2106 cell/ml (Nadeau et al., 1996). Perfusion mode operation has been demonstrated as the most effective measure to control the culture environment and to remove toxic byproducts (Cortin et al., 2004; Henry et al., 2004; Kamen and Henry 2004). Nevertheless, the cell specific productivity could only be maintained by infecting cells at densities up to 3106 cell/ml using high perfusion rate at 2 reactor volumes per day (VVD), at 2 days post-infection, a very costly Importance of cell cycle 109 proposition (Henry et al., 2004; Kamen and Henry 2004). By decreasing the temperature to 35 oC post infection, Cortin et al. (2004) have further increased CCI up to 8106 cell/ml at perfusion rate of 1 VVD with a production of 7.8109 infectious particles/ml (ip/ml). In studies conducted by the authors under much less expensive fed batch mode using 293 cells adapted to non-ammoniagenic medium an 1.8 fold increase on AdV volumetric productivity at CCI 3106 cell/ml was obtained, when comparing with infection at CCI 1106 cell/ml in the same medium, with a production of 8.5109 ip/ml (Ferreira et al., 2005b; Ferreira et al., 2005c). During cell culture processes, especially in batch and fed-batch mode, apart from the continuously changing culture chemical and physical environment, the cells themselves are also continuously varying their physiological and metabolic states. It is well known that during the culturing, the population of cells goes through four growth stages, i.e., lag, exponential, stationary and death-phase, where an individual cell commits to four successive growth phases in the cell cycle, i.e., G1 (gap one phase, between mitosis and DNA synthesis), S (DNA synthesis phase), G2 (gap two phase, between completion of DNA synthesis and mitosis) and M-phase (mitosis phase). In some cases, cells may escape from the cell cycle and enter a resting phase (G0), from where, given suitable condition, they can return to the cell cycle. During the culture process, the cell cycle phase distribution may be continuously changing and no matter in which phase of the cell cycle the cells are infected by AdV, the viral DNA synthesis will only take place 9h after infection (Hodge and Scharff 1969). The effect of cell cycle phase at the time of infection on AdV productivity has not been well described. The scarce reported data suggests that it is both cell line and virus type dependent. For avian AdV, both production of infectious virus and viral DNA synthesis are correlated with events during the S phase of the infected chicken embryo fibroblasts cells (Kraft and Tischer 1978). For HeLa cells, infection with human Ad5 during S phase produced greater yields of E1B 55-kDa-mutant and E4 orf6-mutant viruses Chapter IV.1 110 than did cells infected during G1 or asynchronous cells. However the production of wild Ad and E4 orf3-mutant Ad was not significantly restricted by the cell cycle (Goodrum and Ornelles 1997). In addition, for many other viruses, the production is associated with the S phase of the cell cycle. For example, coxsackievirus production is dependent on G1 or G1/S phase (Feuer et al., 2002); for minute virus of mice, a parvovirus, the conversion of the viral genome from input single-stranded DNA to a doublestranded DNA, the form which undergoes further replication, appears to be an S phase-specific event (Wolter et al., 1980); in the baculovirus-insect cell expression system, the infection yield at G1 or S phase-infection was 1.5-1.8-fold higher than that at G2/M phase-infection (Saito et al., 2002); the replication of the bovine herpesvirus-4 depends on transition through S phase (Vanderplasschen et al., 1995); the herpes simplex virus type 1 vmw65 (VP16) insertion mutant depends on cells infected during S phase for early protein synthesis and replication (Daksis and Preston 1992). Thus the aim of this work is to investigate the importance of the cell cycle phase of the host 293 cells at the time of infection upon AdV productivity, so that more efficient infections at high cell density may be achieved. 2. MATERIALS AND METHODS 2.1. Cell line and culture maintenance Suspension-adapted 293 cells (ATCC-CRL-1573, Rockville, MD) were grown in shake flasks (Corning, NY) in commercially available and chemically defined medium without serum and protein, CD293, supplemented with 4 mM of L-glutamine (all from, Invitrogen, Glasgow, UK) at 160 rpm under an humidified atmosphere of 8% CO2 in air at 37ºC. The cells were routinely propagated twice a week with an inoculum density of 0.5106cell/ml. Importance of cell cycle 111 2.2. Cell infection A replication-defective AdV derived from type 5 AdV was kindly provided by Dr. Tom Barret (IAH-Pirbright, UK). Infection was carried out in 125 ml shake flasks with working volumes of 40 ml using a multiplicity of infection (MOI) of 10 (see Annex A). Infected 293 cells were harvested at 0, 24, 48 and 72 hpi and prepared as previously described in Ferreira et al. (2005b) 2.3. Analytical methods Cell concentration and viability were determined by counting cells on a Fuchs-Rosenthal haemocytometer (Brand, Wertheim, Germany) using the trypan blue (Invitrogen, Glasgow, UK) dye exclusion method. AdV titration was performed by the end-point dilution method (TCID50) using 96 well plates and anchorage-dependent 293 cells cultured in T-flasks (Sarstedt); these cells were cultured in MEM supplemented with heatinactivated 5% fetal bovine serum (all from Invitrogen, Glasgow, UK) and 2 mM L-glutamine under an humidified atmosphere of 5% CO2 in air at 37 oC. The titer was determined according to the method of Spearman and Kraber, as described elsewhere (Darling et al., 1998). Glucose and lactate were analyzed using an YSI Multiparameter Bioanalytical System Model 7100 MBS (Yellow Springs, USA). Ammonia was quantified enzymatically using the UV test number 1112732035 (Roche Diagnostics GmbH, Mannheim, Germany). 2.4. Flow cytometric analysis To quantify the cell-cycle distribution, cells were stained with propidium iodide to measure the DNA content by flow-cytometric analysis (FACS). Cells were centrifuged at 200 g for 10 min at 4ºC, washed with ice-cold PBS Chapter IV.1 112 and stained with ice-cold Vindelov’s solution (Vindelov et al., 1983): 1 g/l trisodic citrate, pH 7.6 (Sigma); 50 mg/l propidium iodide (Sigma); 0.1% NP 40 (Roche); 700 U/l RNase A (Sigma); 0.01 M NaCl (Merck) for a final cell concentration of 1.5106cell/ml. All FACS analyses were performed on a FACSCaliburTM using the Cell QuestTM software (Beckton-Dickinson, San Jose, CA). 3. RESULTS AND DISCUSSION 3.1. Relationship between cell density and cell cycle phase distribution during normal batch cultures of 293 cells 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 Time after inoculation (days) Viable cell density 0 10 20 30 40 50 60 70 Phase in the cell cycle (percentage of population) 0 1 2 3 4 5 6 7 8 0 2 3 4 5 6 0 10 20 30 40 50 60 70 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 Viable cell densityViable cell density (106cell/ml) 0 10 20 30 40 50 60 70 Phase in the cell cycle (percentage of population) Phase in the cell cycle (percentage of population) 0 1 2 3 4 5 6 7 8 0 2 3 4 5 6 0 10 20 30 40 50 60 70 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 Time after inoculation (days) Viable cell density 0 10 20 30 40 50 60 70 Phase in the cell cycle (percentage of population) Phase in the cell cycle (percentage of population) 0 1 2 3 4 5 6 7 8 0 2 3 4 5 6 0 10 20 30 40 50 60 70 0 1 2 3 4 5 6 7 8 0 1 2 3 4 5 6 Viable cell densityViable cell density (106cell/ml) 0 10 20 30 40 50 60 70 Phase in the cell cycle (percentage of population) Phase in the cell cycle (percentage of population) 0 1 2 3 4 5 6 7 8 0 2 3 4 5 6 0 10 20 30 40 50 60 70 Figure 4.1.1. Profiles of cell growth and cell cycle phase distribution during 293 cell suspension batch cultures. Symbols: , cell density; , G2/M phase; , G1/G0 phase; ▲, S phase. Figure 4.1.1 describes the cell cycle phase distribution at different cell densities during a normal batch cultivation process of 293 cells. After inoculation, the proportion of cells in the S phase of the cell cycle Importance of cell cycle 113 increased sharply, with a 70% increase within the first 24 h, from ca. 30% to ca. 50% of the total cell population. Conversely, the percentage of G1/G0 cells in the cell culture fell during this period. The highest S phase proportion occurred at the middle of the lag phase of the growth stage, corresponding to a cell density of 0.6-0.8106cell/ml. Thereafter, the percentage of cells in the S phase decreased gradually, as the cell density increased above 1106cell/ml; correspondingly, the percentage of G1/G0 cells in the cell culture rose, while the percentage of cells at G2/M phase was kept relatively stable at about 11% during the whole growth cycle. Accompanying the cell density increase, the nutrients in the culture medium, namely glucose and glutamine, decreased quickly; at the same time, the accumulation of the byproducts, specially lactate and ammonia, increased rapidly in the 293 cell suspension batch cultures (Figure 4.1.2). 0 5 10 15 20 25 30 0 1 2 3 4 5 6 Time after inoculation (days) Concentration of Glc and Lac (mM) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Concentration of Gln and NH 4 +(mM) 0 5 10 15 20 25 30 0 1 2 3 4 5 6 Time after inoculation (days) Concentration of Glc and Lac (mM) 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 Concentration of Gln and NH 4 +(mM) Figure 4.1.2. Kinetics of total glucose (Glc, ) and glutamine (Gln, ) consumption and total lactate (Lac, ) and ammonia (NH4+, ∆) production of 293 cell suspension batch cultures. Although Lullau et al. (2003) observed for 293-BACE-Fc a higher S phase Chapter IV.1 114 fraction at the end of batch culture, the trend in the cell cycle phase distribution of 293 cells obtained here is well supported by others who observed similar results for cultures of hybridoma cells (Ramirez and Mutharasan 1992; Balcarcel and Stephanopoulos 2001; Luo and Yang 2004), CHO cells (Ley and Tobey 1970; Tobey and Ley 1970) and Hela cells (Koza and Herbst 1992; Goodrum and Ornelles 1997). 3.2. Effect of different proportion of cells under S phase in equalized cell density conditions upon AdV productivity As demonstrated above, the proportion of cells in the S phase decreased as the cell density rose. To evaluate if this decreased proportion of cells in the S phase at high cell density correlates with the so called “cell density effect”, cells were infected at different growth stages with different proportions at S phase under equalized cell density conditions at CCI of 1106 cell/ml. In brief, after harvesting, the cells were centrifuged to totally remove the spent medium, and then resuspended in fresh medium at a final concentration of 1106 cell/ml before infection, and infected with the same MOI at the same time. As stated in the Introduction, the CCI of 1106 cell/ml was chosen on the basis of the observations that it is the optimum CCI for normal AdV production (Ferreira et al., 2005c). The medium exchange was performed to avoid lack of nutrients and presence of toxic byproduct thus permitting to isolate the evaluation of the cell cycle effect. The results are illustrated in Figure 4.1.3 and summarized in Table 4.1.1. Importance of cell cycle 115 1E+00 1E+01 1E+02 1E+03 1E+04 0 24 48 72 Time post infection (hours) ip/cell 0 24 48 72 1E+00 1E+01 1E+02 1E+03 1E+04 0 24 48 72 Time post infection (hours) ip/cell 0 24 48 72 1E+00 1E+01 1E+02 1E+03 1E+04 0 24 48 72 Time post infection (hours) ip/cell 0 24 48 72 1E+00 1E+01 1E+02 1E+03 1E+04 0 24 48 72 Time post infection (hours) ip/cell 0 24 48 72 Figure 4.1.3. Profiles of cell specific productivity of cells obtained from different growth stages with different proportions in S phase at the time of infection in equalized cell density conditions at a CCI of 1106 cell/ml. The 293 cells for this experiment were harvested from lag, middle exponential, late exponential and stationary phases, corresponding to cell densities of 0.7, 1.6, 3.3, and 5106 cell/ml, with proportions of cells in S phase of 50% (▲), 40% (), 32% () and 28% (Δ), respectively. The cells were centrifuged to discard the spent medium and replaced with fresh medium at CCI of 1106 cell/ml, in order to provide an equalized infection environmental condition for all the cells. Table 4.1.1. Effect of cells obtained from different growth stages with different proportions in S phase at the time of infection in cell specific productivity at the optimal cell concentration at infection (1106 cell/ml). Infections were performed at MOI of 10 and AdV harvested at 72 hpi. Growth stage of cells Original cell density (cell/ml) Percentage of S phase cells at infection Maximum cell specific productivity (ip/cell) Lag 0.710650%  3% 8.3103 0.7103 Middle exponential 1.610640%  3% 4.7103 0.4103 Later exponential 3.310632%  2% 2.5103 0.3103 Stationary 5.0 28%  2% 1.3103 0.5103 Chapter IV.1 116 The infection of cells from the lag phase, with the highest proportion at S phase, resulted in the highest productivity of 8.3103ip/cell at 72 hpi, 6.4 times higher than the level of cells obtained from stationary phase with the lowest proportion in S phase at the same time (1.3103ip/cell). In other words, even though all infections were performed at the optimal CCI (1106 cell/ml), cells obtained from different growth stages exhibited different cell specific productivities. These results strongly support the hypothesis that the cell specific productivity is dependent upon cell cycle at infection, i.e., higher proportion of cells at S phase, corresponding to a larger cell population under faster division, resulted in higher AdV productivities. Interestingly, cells obtained from the growth stages at higher cell densities of 3.3 and 5.0106 cell/ml presented a lower cell specific production rate and a higher optimal harvesting time of 72 hpi although 48 hpi is the optimal harvesting time for the cells obtained from the growth stages at lower cell densities of 0.7 and 1.6106 cell/ml (Figure 4.1.3). Although this effect could be correlated with cell aggregation in the experiments performed at high cell densities, no significant cell aggregation (no more than 2-4 cells aggregates) was found. Moreover, an increase of 60% on viable cell density at 24 hpi for cells with the higher proportion in S phase was observed (Figure 4.1.4). Hodge and Scharff (1969) observed that no matter in which phase of the cell cycle the cells are infected, the viral DNA will start to be produced at 9 hpi, meaning that a cell in G1 phase will stop at the S phase and that a cell in the S phase will double once more after infection; therefore it is not surprising to observe cell growth after infection where the proportion of cells on the S phase of the cell cycle is higher. Importance of cell cycle 117 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 0 24 48 72 Time post infection (hours) Viable cell density (106cell/ml) 0 24 48 720 24 48 72 Time post infection (hours) Viable cell density (106cell/ml) 0 24 48 72 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 0 24 48 72 Time post infection (hours) Viable cell density (106cell/ml) 0 24 48 720 24 48 72 Time post infection (hours) Viable cell density (106cell/ml) 0 24 48 72 Figure 4.1.4. Profiles of cell growth after infection in equalized cell density conditions at a CCI of 1106 cell/ml. The 293 cells for this experiment were harvested from lag, middle exponential, late exponential and stationary phases, corresponding to cell densities of 0.7, 1.6, 3.3, and 5106 cell/ml, with proportions of cells in S phase of 50% (▲), 40% (), 32% () and 28% (Δ), respectively. The effect of different proportion of cells under S phase at infection on AdV productivity was further investigated at higher CCI in the equalized cell density conditions. In this case, the CCI of 4106 cell/ml was chosen in order to highlight the difference in cell specific productivity in the equalized cell density condition. As illustrated in Figure 4.1.5, similarly to the results achieved at the optimal CCI of 1106 cell/ml, at the CCI of 4106 cell/ml the infection of cell cultures with higher proportion in S phase also resulted in higher cell specific productivity. At this CCI, the infection of cells with the highest S phase proportion (50%), resulted in a production of 2.5103 ip/cell at 72 hpi. This cell specific productivity was 16.7-fold higher than that obtained from infection of cells from the stationary growth stage at a cell density of 5106 cell/ml, consisting of the lowest S phase cell proportion (28%), and resulted in the higher volumetric productivity of 11010 ip/ml (Figure 4.1.5). Once more, this constitute strong evidence that the proportion of cells in S phase at infection is an Effect of cell cycle synchronisation on AdV production 125 PART 2 293 cell cycle synchronisation in adenovirus production Ferreira T.B., Perdigão R., Carina A.C., Zhang C., Aunins J.G., Carrondo M.J.T., Alves P.M. Biotechnol. Bioeng. (submitted) Chapter IV.2 126 Abstract As the market requirements for adenovirus vectors (AdV) increase, the maximisation of the virus productivity per culture volume per unit time is a key requirement. However, despite the fact that 293 cells can grow up to 8×106 cell/ml in simple batch mode operations, for optimal AdV infection a maximum cell density of 1×106 cell/ml at infection time has been reported. In addition, AdV production appears to be dependent of the cell cycle phase at the time of infection. To evaluate the dependence of AdV production on cell cycle phase, 293 cells were chemically synchronised at each phase of the cell cycle; a 2.6 fold increase on cell specific productivity was obtained when the percentage of cells at the S phase of the cell cycle was increased from 36% to 47%. A mathematical equation was used to relate AdV productivities with S phase cell synchronisation using this data. A temperature shift strategy was also used attempted for synchronisation at the S phase, avoiding the use of chemical inhibitors. S phase synchronisation was possible by decreasing the culture temperature to 31ºC during 67h and restoring it to 37ºC during 72h. By using this strategy we were able to synchronise 57% of the population in the S phase of the cell cycle obtaining an increase of 7.3 fold on cell specific productivity after infection. Effect of cell cycle synchronisation on AdV production 127 CONTENTS 1. INTRODUCTION________________________________________ 128 2. MATERIALS AND METHODS _______________________________ 130 2.1. Cell line and medium _______________________________ 130 2.2. Adenoviral Vector__________________________________ 131 2.3. Chemical synchronisation of 293 cells __________________ 131 2.4. Infection of chemically synchronised 293 cells____________ 132 2.5. Temperature Synchronisation of 293 cells _______________ 133 2.6. Infection of 293 cells synchronised by temperature shift____ 133 2.7. Flow cytometric analysis_____________________________ 134 2.8. Virus culture samples preparation _____________________ 134 2.9. Adenoviral vector titration ___________________________ 134 2.10. Cell concentration and viability determination __________ 135 3. RESULTS AND DISCUSSION _______________________________ 135 3.1. 293 cell synchronisation with chemical compounds________ 135 3.2. Adenovirus vector production in chemically synchronised 293 cells ________________________________________________ 137 3.3. Model for adenovirus production in chemically synchronised cells ____________________________________________________ 139 3.4. 293 cell synchronisation by temperature shift____________ 141 3.5. Adenovirus vector production in synchronised 293 cells by temperature shift______________________________________ 144 4. CONCLUSIONS_________________________________________ 145 5. ACKNOWLEDGEMENT ___________________________________ 146 6. REFERENCES__________________________________________ 147 Chapter IV.2 128 1. INTRODUCTION Adenoviruses are leading vectors for gene transfer, with products in development for applications ranging from cancer therapeutics to prophylactic vaccines and replacement therapies for genetic deficiencies (for reviews see Gallo et al. (2005), Ferreira et al. (2005a) and Morsy et al., (1998)); worldwide, over 300 clinical trials using Adenovirus Vectors (AdV) have already been approved, ongoing or completed, making AdV the most used delivery vector in clinical trials (www.wiley.co.uk/genetherapy/ clinical). One of the goals of a commercial AdV cultivation process is to maximize the virus productivity per culture volume per unit time. The cell density at infection is a very important parameter as it impacts on the AdV volumetric productivity. However, despite the fact that 293 cells can grow up to 8×106 cell/ml in simple batch mode operations, a maximum infection cell density of 1×106 cell/ml has been reported as optimal (Nadeau and Kamen, 2003; Xie et al., 2003; Kamen and Henry, 2004; Ferreira et al., 2005c; Maranga et al., 2005). An approach to maintain cell specific productivity at increased cell density consists of infecting the cells after medium exchange (Garnier et al., 1994; Iyer et al., 1999; Ferreira et al., 2005b). However, this procedure adds the extra complexity of a cell separation step and increases in product production cost due to medium exchange. Other approach consists in the use of a fed-batch strategy, but its success has been limited (Nadeau et al., 1996; Ferreira et al., 2005b). By using a non-ammoniagenic medium an increase of 5 fold on cell specific productivity could be obtained at a CCI of 3×106 cell/ml still bellow the expected (Ferreira et al., 2005b). AdV production is usually performed assuming that cultured cells represent a uniform target for the virus. However, in fact, cells in culture are not homogenous, including a collection of cells with distinct physiologies. A major determinant of cell physiology in culture is the cell cycle with its Effect of cell cycle synchronisation on AdV production 129 four stages (Stein et al., 1999) including elevated protein synthesis (G1 and G2 phases), elevated DNA synthesis (S phase) and changes in membrane trafficking, cytoskeletal organization and cell-cell interactions (M phase). In Chapter IV, Part 1 was shown that AdV production seems to be dependent of the cell cycle phase at the time of infection: during cell growth the highest percentage of cells at the S phase of the cell cycle is present at approximately 1×106 cell/ml, which has been shown to be the optimal concentration for AdV infection, decreasing with increasing cell density (Zhang et al., 2006). Furthermore, Goodrum and Ornelles (1997) observed that in synchronously growing HeLa cells, approximately 75% of the cells infected during S phase, produced E1B AdV mutant virus, whereas only 10% of the cells infected during G1 phase were able to produce virus. Therefore, it is pertinent to synchronize cells at each phase of the cell cycle in order to identify the best phase for infection to increase cell specific productivity at high cell densities. Several methods for obtaining synchronised populations of mammalian cells in vitro have been reported by using genetic, chemical, environmental and physical strategies (Davis et al., 2001). For reversible chemical synchronisation several compounds are available, as thymidine (Law et al., 2006) for G1 phase synchronisation, mimosine (Hughes and Cook, 1996) for late G1 phase synchronisation, aphidicolin (Jackson, 1995), rapamicine (Law et al., 2006) and hydroxyurea (Rosenkranz and Becker, 1973) for synchronisation at the G1/S transition phase, staurosporine (Kocher and Clemetson, 1991) for G2 phase synchronisation and nocodazole (Ho et al., 2001) for synchronisation at the G2/M transition phase. However, in addition to the difficulty in ensuring their elimination in the final product, the use of these chemicals as synchronising agents also affects cell viability, often presenting severe limitations for large scale production purposes. Besides chemical synchronisation, several studies suggest hypothermia as a potentially useful tool to manipulate mammalian cell cycle distribution (Rao and Engelberg, 1965; Sisken et al., 1965; Watanabe and Okada, 1967). Chapter IV.2 130 Typically, mammalian cells are grown at 37ºC, mimicking mammalian body temperature. At “sub-optimal” temperatures (25-33ºC) there is a prolongation of the cell cycle duration as a result of the temporal dilation of the G1, S and M (mitosis) phases, with G2 being the least and M the most sensitive phases to temperature shift (Rieder and Cole, 2002), undergoing the major increase in duration (Rao and Engelberg, 1965; Sisken et al., 1965). However, theses occurrences are very dependent on the cell line considered. For example, incubation of human diploid fibroblasts (Enninga et al., 1984) or of a mouse leukemic cell line (Watanabe and Okada, 1967) at 31ºC induced cell cycle arrest in G1/G0 phase; on the other hand, when HeLa cells were incubated at sub-normal temperatures there was an accumulation of cells in the M phase (mitosis) (Rao and Engelberg, 1965). In this work, we report on the effect of the cell cycle on AdV production. First, in order to establish the best cell cycle phase for AdV infection, 293 cells were chemically synchronised. Then, due to the inherent problems associated with the use of these chemicals compounds for production purposes, a synchronisation strategy based on temperature shift was developed and evaluated in terms of AdV production at high cell densities in 2 L bioreactors. 2. MATERIALS AND METHODS 2.1. Cell line and medium Anchorage-dependent 293 cells, purchased from Stratagene (Catalog #240085), were cultured in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% (v/v) heat-inactivated (56 ºC, 30 min.) Foetal Bovine Serum (FBS) (all from Invitrogen, Glasgow, UK) and 4.5 g/L of glucose (Merck, Darmstadt, Germany) at a humidified atmosphere of 5% CO2 in air at 37ºC. 293 cells were routinely propagated twice a week after trypsin/EDTA (Invitrogen, Glasgow, UK) addition at room temperature. Effect of cell cycle synchronisation on AdV production 131 Suspension adapted 293 cells were grown in CD 293 (Invitrogen, Glasgow, UK), supplemented with 4 mM of glutamate (non-ammoniagenic medium) (Sigma-Aldrich, St. Louis, MO) at a humidified atmosphere of 8% CO2 in air at 37ºC in shake flasks (Corning, NY). Cells were routinely propagated twice a week using an inoculum of 0.5×106 cell/ml. 2.2. Adenoviral Vector Recombinant AdV expressing the GFP protein were generated by homologous recombination of plasmid and digested DNA in E. coli strain BJ5183 (rec+). The GFP shuttle vector was first cloned into a transfer vector; the resulting plasmid was then linearized with a restriction enzyme (Swa I) and co-transformed into E. coli together with the plasmid containing AdV (pKP1.3). Recombinants were selected with ampicillin and screened by restriction enzyme analysis. The recombinant AdV construct was subsequently cleaved with Pac I to expose its ITR (Inverted Terminal Repeats) and transfected into 293 cells to produce viral particles. 2.3. Chemical synchronisation of 293 cells For chemical synchronisation, 293 cells were seeded in 6 well plates at 1106 cell/well, with a culture volume of 2 mL of DMEM supplemented with 10% (v/v) FBS and 1% (v/v) antibiotic solution (10000 µg/mL streptomycin and 10000 U/mL penicillin) (Sigma-Aldrich, St. Louis, MO). One day after, seeded cells were incubated with each inhibitor at three different concentrations during 6, 12, 24 and 48h. A control for each inhibitor where only the respective solvent was added to the cell culture was performed. The experimental design is summarised in Table 4.2.1. After the incubation period, 293 cells were digested at room temperature with trypsin/EDTA and analysed by flow cytometry. Chapter IV.2 132 Table 4.2.1. Reversible chemical inhibitors used and correspondent cell cycle arrest point. Stock solutions, tested concentrations and control experiments are shown for each case. Inhibitor Type Target Arrest Point Inhibitor Stock solution Tested concentrations Control Mimosine Late G1 100 mM in a PBS solution 0.1% (v/v) in DMSO 1 mM 1 M 1 nM PBS 0.1% (v/v) in DMSO Hydroxyurea G1/S transition 100 mM in a PBS solution 0.1% (v/v) in DMSO 1 mM 1 M 1 nM PBS 0.1% (v/v) in DMSO Aphidicolin G1/S transition 2.96 mM in DMSO 0.1 mM 1 M 1 nM DMSO Staurosporine G2 0.214 mM in ethanol 96% (v/v) 10 mM 1 M 0.1 M Ethanol 96% (v/v) Nocodazole G2/M transition 13.3 mM in a PBS solution 0.1% (v/v) in DMSO 1 mM 1 M 1 nM PBS 0.1% (v/v) in DMSO 2.4. Infection of chemically synchronised 293 cells Infection of chemically synchronised cells was performed by seeding 293 cells in 6 well plates at 1106 cell/well, with a culture volume of 2 mL of DMEM supplemented with 10% (v/v) FBS and 1% (v/v) antibiotic solution (10000 µg/mL streptomycin and 10000 U/mL penicillin). One day after, the best inhibitory conditions, determined beforehand, were added: 1 mM hydroxyurea during 48h, 0.1 mM aphidicolin during 12h and 1 µM staurosporine during 48h. In the experimental controls only the solvents were added. After the incubation period, cells were centrifuged at 200 g during 10 min. at 4ºC and suspended in 500 µL of fresh DMEM without FBS Effect of cell cycle synchronisation on AdV production 133 and reseeded in a new well. Then, cells were infected with AdV using a multiplicity of infection (MOI) of 10. 30 min. afterwards, 1 mL of DMEM supplemented with 10% (v/v) FBS and 1% (v/v) antibiotic solution (10000 µg/mL streptomycin and 10000 U/mL penicillin) was added to each well. Infected 293 cells were then harvested at several times after digestion at room temperature with trypsin/EDTA. 2.5. Temperature Synchronisation of 293 cells For temperature synchronisation, 293 cells adapted to grow in suspension were inoculated at 0.5×106 cell/ml in 125 mL shake flasks with a working volume of 40 mL in a humidified atmosphere of 8% CO2 in air at 37ºC. Once the concentration of 2×106 cell/ml cells was reached, incubation at 29, 31, 33, 35 e 37ºC was carried out for 72h, and the cell cycle followed by flow cytometry. 2.6. Infection of 293 cells synchronised by temperature shift Infection of 293 cells synchronised by temperature shift was done in 2 L bioreactors (Braun, Melsungen, Germany). For this purpose, two bioreactors were inoculated at 0.5×106 cell/ml with 293 cells adapted to grow in suspension. Once the population reached the concentration of 3×106 cell/ml, one of the bioreactors (control) was infected and the other was synchronised at the S phase of the cell cycle and infected. All the infections were done using a MOI of 10 and AdV harvested at 48 hpi. Synchronisation was performed by firstly lowering the culture temperature to 31ºC during 67h and then restoring it to 37ºC during 72h. The agitation rate was maintained at 110 rpm (NRe = 6800); pH was controlled at 7.2 by aeration with a CO2 gas-mixture and NaOH 0.2 M; the dissolved oxygen was controlled at 80% air saturation.