Eric Becker

An XBP-1 dependent bottle-neck in production of IgG subtype antibodies in chemically defined serum-free Chinese hamster ovary (CHO) fed-batch processes

The optimization of production processes for therapeutic antibodies is a continuing challenge in pharmaceutical biotechnology. Although it could be demonstrated that vector design and host cell engineering can improve transcriptional and translational efficiency and thereby result in generation of high producer cell lines, it is not clear whether introduction of transgenes that regulate protein transport or affect post-translational modifications could further improve such industrial processes. Here, we show that heterologous expression of the transcription factor X-box binding protein-1 (XBP-1) can lead to an increase in endoplasmic reticulum (ER) content and specific therapeutic antibody productivity of Chinese hamster ovary (CHO)-DG44 cells in inoculum suspension cultures. This effect translates into 40% increased overall antibody titers in a fed-batch format where cells are grown in chemically defined serum-free media. Protein-A purified antibody products from mock-transfected cells and XBP-1 transfected cells were found to be of comparable quality with regard to glycosylation pattern and physicochemical characteristics. The data demonstrate the potential of XBP-1 engineering to improve mammalian cell culture production processes to yield high amounts of a therapeutic protein product of desired quality.

Heterologous expression of the lipid transfer protein CERT increases therapeutic protein productivity of mammalian cells.

Recent studies have demonstrated that the introduction of transgenes regulating protein transport or affecting post-translational modifications can further improve industrial processes for the production of therapeutic proteins in mammalian cells. Our study on improving therapeutic protein production in CHO cells by heterologous expression of the ceramide transfer protein (CERT) was initiated by the recent discovery that CERT is involved in protein kinase D (PKD)-dependent protein transport from the Golgi to the plasma membrane. We generated a set of CHO DG44 cell lines by stable integration of constructs expressing either CERT wild-type or CERT S132A, a mutant conferring increased lipid transfer activity, or a mock plasmid. CHO cells expressing heterologous CERT demonstrated significantly higher specific productivities of the therapeutic protein HSA when grown in inoculum suspension cultures. This effect translated into significantly increased overall HSA titers in a fed-batch format where cells are grown in chemically defined serum-free media. Furthermore, we could show that CERT also enhanced monoclonal antibody secretion in two IgG production cell lines with different basal productivities. The data demonstrate the potential of CERT engineering to improve mammalian cell culture production processes to yield high amounts of a therapeutic protein product of desired quality. To our knowledge, this is the first study showing a bottle neck in recombinant protein secretion at the Golgi complex in mammalian cells.

Evaluation of a combinatorial cell engineering approach to overcome apoptotic effects in XBP-1(s) expressing cells

Genetic engineering of producer cell lines for production of therapeutic antibodies in order to increase the yield of production processes remains a continuing challenge. Recently it was shown that heterologous expression of the active, spliced form of human X-box binding protein 1 (XBP-1(s)) can increase the amount of secreted protein products in mammalian cell culture processes. However, a prerequisite for the industrial application of any cell engineering approach is the ability to generate monoclonal cell lines that stably express the engineering gene to maintain the desired phenotype. Here, we show a decrease in heterologous human XBP-1(s) expression in CHO production cells producing a therapeutic antibody product monitored over a prolonged period in serial culture. Colony formation assays (CFA) in CHO-K1 cells reveal a general survival disadvantage conferred by XBP-1(s) in this cell type. We aimed to rescue this phenotype by expressing the caspase-inhibitor XIAP (x-linked inhibitor of apoptosis). Using a set of bicistronic expression vectors we engineered an antibody producing CHO cell line with XBP-1(s) and XIAP alone and in combination. Interestingly, co-expression of both genes resulted in the highest specific productivities (Qp) and final titers in a serum-free fed-batch process in chemically defined media. Thus, the combination of secretion and anti-apoptotic engineering provides an interesting approach for future applications in industrial mammalian cell culture.

Supplementation of serum free media with HT is not sufficient to restore growth properties of DHFR-/- cells in fed-batch processes - Implications for designing novel CHO-based expression platforms.

DHFR-deficient CHO cells are the most commonly used host cells in the biopharmaceutical industry and over the years, individual substrains have evolved, some have been engineered with improved properties and platform technologies have been designed around them. Unexpectedly, we have observed that different DHFR-deficient CHO cells show only poor growth in fed-batch cultures even in HT supplemented medium, whereas antibody producer cells derived from these hosts achieved least 2-3 fold higher peak cell densities. Using a set of different expression vectors, we were able to show that this impaired growth performance was not due to the selection procedure possibly favouring fast growing clones, but a direct consequence of DHFR deficiency. Re-introduction of the DHFR gene reproducibly restored the growth phenotype to the level of wild-type CHO cells or even beyond which seemed to be dose-dependent. The requirement for a functional DHFR gene to achieve optimal growth under production conditions has direct implications for cell line generation since it suggests that changing to a selection system other than DHFR would require another CHO host which - especially for transgenic CHO strains and tailor-suited process platforms - this could mean significant investments and potential changes in product quality. In these cases, DHFR engineering of the current CHO-DG44 or DuxB11-based host could be an attractive alternative.