Gene transfer and gene amplification in mammalian cells

Florian M. Wurm and Martin Jordan

1. Introduction on the origin of Chinese hamster ovary cells for recombinant
protein production
Mammalian genomes have exceptional plasticity. They can exhibit short-term
plasticity as evidenced by the fact that some DNA sequences, whether endogenous or
introduced by artificial gene transfer, can be induced and/or selected for gene
amplification to high copy numbers. In addition, mammalian genomes have an
apparently unlimited capacity to take up foreign DNA, the hallmark of this being
the presence of retroviral and bacterial DNA sequences that have become an integral
part of these genomes. The short-term plasticity and ability to take up DNA
makes cell lines useful for the expression and production of recombinant proteins of
scientific and pharmaceutical value. Foreign genes can be easily integrated into
mammalian cells cultivated in vitro and, through selection, populations of cells of
clonal origin can be identified that show an increased copy number of the inserted
genes. These cell lines express the required genes of interest at elevated levels.
This approach has been used widely in both industry and academia, particularly with
two immortalized cell lines, Chinese hamster ovary (CHO) cells and a mouse
myeloma cell line, NS0. We will primarily concentrate on the dhfr/CHO system,
since it is the most commonly used system. In general, most of the observations
made with this system apply to other immortalized cell lines and thus should give
insights into other types of gene transfer and amplification systems in mammalian
cells.
CHO cells remain the most popular mammalian host for recombinant protein
production since the mid 1980s, when the first product, human recombinant tissuetype
plasminogen activator Activase_ (rTPA), received regulatory approval for
marketing. These cells have also been a preferred choice for studies on gene transfer
and gene amplification in mammalian cells. Some of the studies were driven by
questions raised by regulatory authorities controlling the clinical evaluation and
marketing of pharmaceutical products. Other motivations were the need to maximize
S.C. Makrides (Ed.) Gene Transfer and Expression in Mammalian Cells
_ 2003 Elsevier Science B.V. All rights reserved

gene expression from transfected and amplified plasmid sequences and to establish
recombinant cell lines with long-term stability with regard to the genome and to
protein productivity. The popularity of CHO cells in recombinant protein production
is based on both practical and virological safety reasons. Discussions on virus
safety with mammalian production host systems started a half century ago when the
first biological products, i.e., vaccines and natural interferons, were developed using
primary and transformed mammalian cells. It is an essential goal both for the
manufacturers of recombinant protein pharmaceuticals and for regulatory agencies
to minimize any potential risks associated with the use of immortalized recombinant
mammalian cell hosts.
CHO cells were established for the study of somatic cell genetics in 1957 by Puck
[for review see [1]]. These cells are well characterized with respect to karyotype,
chromosome structure, gene mapping and culture conditions. Upon the emergence
of DNA technology one particular derivative of the original CHO cells (see below)
was used in a number of pioneering gene transfer experiments [2,3]. It was possible to
sub-cultivate these cells after the identification of genetically altered candidate clones
producing the protein(s) of interest under increasingly selective conditions. This
resulted in the generation of more productive amplified subclones. Another key
feature for industrial scale production was the finding that CHO cells can be easily
adapted to grow in suspension. In the mid 1980s R. Arathoon, A. Lubiniecki, G.
Polastri and others at Genentech Inc. in San Francisco developed a 10,000-liter
production process for rTPA, using deep tank technology with suspensionadapted

CHO cells [4].

CHO cells are considered to be a safe host for the production of therapeutic
proteins, which will be administered parenterally to human patients. Human
pathogenic viruses like Polio, Herpes, Hepatitis B, HIV, Measles, Adenoviruses,
Rubella and Influenza do not replicate in these cells. Thus, the risk of a viral
adventitious agent being involuntarily carried along with the product of interest
is low. Wiebe et al. tested a total of 44 human pathogenic viruses for replication
in CHO cells and found that only seven of these (Reo 1,2,3, Mumps, and
Parainfluenza 1,2,3) were able to infect these cells [5]. These safety considerations
were discussed extensively among scientists from both regulatory agencies and
the bio-pharmaceutical sector. Today one can only speculate whether an
immortal human cell substrate with replicative capacity for a larger number of
human viruses would have received approval for the production and marketing
of a protein therapeutic. Only very recently (2002) was the marketing approval
given for a protein pharmaceutical produced in the immortalized human
cell line HEK-293 (Activated Protein C, requiring C-terminal carboxylation,
E. Lilly). Due to complex processing and post-translational modifications,
many therapeutic proteins must be produced in mammalian cells. At present
(2003), at least 50 recombinant proteins made in CHO cells have received
approval for marketing. It is difficult to assess the exact number of CHO-based
product candidates that are in pre- and clinical evaluation, but it is with certainty
above 200.