Search

Background CHO (Chinese Hamster Ovary) cells are the cell line of choice for therapeutic protein production. Although the achieved volumetric titers have increased significantly over the past two decades, the establishment of wellproducing CHO cell lines is still difficult and not always successful [1]. Factors influencing productivity are the chosen host cell line, the genetic vectors, applied media, the cultivation strategy as well as the product itself. Several CHO host strains are available for recombinant protein production, however, they are often quite diverse in terms of growth rate, maximal achieved cell concentrations and specific productivities. Specific productivity is also related to the locus of integration of the transgenes due to positional effects caused by the chromatin environment. Previously it was described that Bacterial Artificial Chromosomes (BACs) carrying the Rosa26 locus are advantageous for the recombinant protein production in CHO cells, enhancing the specific productivity compared to plasmid derived recombinant CHO cells [2-4]. In this project we aim to identify factors influencing volumetric productivity using different CHO hosts, Rosa 26 BACs as genetic constructs and suitable cell culture media. First, different commonly used CHO host cell lines were analyzed in various cell culture media to identify which host strain performs best. Secondly, we generated a recombinant cell line, producing the highly glycosylated HIV envelope protein gp140 as an example for a difficult to express model protein. Gp140 expression was compared to an already existing gp140 cell line generated by a plasmid vector as expression system. Methods Cell culture: CHO-DUKX-B11 (ATCC-CRL-9096) and CHO-DG44 (life technologies) were serum-free cultivated in spinner flasks. CHO-K1 (ATCC-CCL-61) and CHO-S (life technologies) were serum-free cultivated in in shaker flasks. BAC Recombineering: E.coli carrying the Rosa 26 BAC (~220 kbp) were transformed with a plasmid coding for a recombinase. Consecutively, a plasmid carrying the gp140 (CN54) gene flanked by homologous regions to the BAC was used for the transformation of the recombinase positive E.coli cells. BAC positive colonies were selected and the BAC DNA was purified (NucleoBond Xtra BAC, Macherey Nagel). Transfection and selection: CHO-S host cells were transfected with linearized, lipid complexed (Lipofectin) CN54 Rosa26 BAC DNA. Recombinant clone selection was performed in 96-well plates using 0.5 mg/mL G418. BAC transfected CHO cells are able to express the transgene as well as a Neomycin resistance gene within the Rosa26 locus.

Antibody engineering allows proper design of improved properties in antibody products such as a high binding affinity, stable biophysical properties and low immunogenicity. Besides these obvious features improved by elegant design the primary amino acid sequence also has tremendous effects on the performance of the expression host itself influencing cell culture parameters including specific productivities and growth rates therefore having major impact on the finally obtained antibody expression titers [1]. By understanding the underlying mechanisms how certain primary sequence combinations influence culture performance in conjunction with biophysical features of the antibody protein, we might rationally improve the production process of these expensive but essential biotherapeutic products. Many different antibody variants were expressed in our group in different projects based on various isotypes. For several antibody variable regions we could observe a direct correlation of advantageous cell performance and a high degree of similarity to the closest human germline sequence of the respective mature antibody leading to higher specific productivities. However, the fundamental principles of these consistent observations and the correlation of cell behavior and biophysical product properties remain quite elusive. Therefore, in this project we aim to set the proper foundations to investigate the generalization of a concept to improve cell performance and antibody expression based on the primary sequence at a high germinality degree (Figure ​(Figure1).1). To have proper control of transgene integration locus, gene copy numbers and mRNA transcript level a suitable host for recombinase-mediated cassette exchange (RMCE) was developed to compare different antibody variants under isogenic conditions. Open in a separate window Figure 1 Observed correlation of specific productivity and sequence identity to the closest human germline sequence (germinality) (A). Antibody expression levels are influenced by several factors. Recombinase-mediated cassette exchange was applied to control for gene copy number and mRNA transcript level. Low expressing antibody variants (2F5, 4B3 and 2G12) were compared to its closest human germline sequence together with a therapeutic antibody (Ustekinumab) with low degree of germinality (B). Respective germline antibody variants were constructed by combination of germline V-, D- and J-segments.

Immunoglobulin A (IgA) mediates a key role in mucosal immunity and is a promising novel immunotherapeutic candidate. However, difficulties in obtaining enough material often hamper in vivo explorations. We have previously generated recombinant Chinese hamster ovary (CHO) cell lines which expressed two different HIV-1 antibodies, 3D6 and 4B3, as IgA1 [1]. One cell line (3D6-IgA) shows high production rates, whereas the other (4B3-IgA) secretes rather low amounts of product. In order to unravel the mystery of productivity bottlenecks we extensively characterized the cell lines regarding growth rate, IgA productivity in long-term culture, immunofluorescence microscopy, flow cytometry and Western blotting of intra- and extracellular product (data not shown). The generated data encouraged us to analyze whether the observed antibody productivities could be explained by gene copy number (GCN) or mRNA levels.

Background Sexually transmitted diseases are predominantly acquired via mucosal membranes of the rectal or genital tract during sexual intercourse. This major port of virus entry is naturally defended by the humoral immune response, with immunoglobulin A (IgA) as the primary antibody class to elicit mucosal immunity. Dimeric IgA (dIgA) reaches the luminal side of mucosal tissues by transcytosis through epithelial cells lining the mucosa. In a first step, dIgAs specifically bind to the basolaterally expressed polymeric immunoglobulin receptor (pIgR) on epithelial cells. For release of IgAs on the luminal side the extracellular portion, termed secretory component (SC), remains attached to the antibody to form secretory IgA (sIgA) [1,2]. One example of a sexually transmitted disease is the human immunodeficiency virus (HIV) which annually infects several million individuals on a global scale and potentially leads to the acquired immunodeficiency syndrome (AIDS). Although current therapies can reduce disease progression in infected individuals, no cure is yet available or within reach in near future. As a consequence increased attention is now being paid to develop drugs that could prevent virus acquisition. 3D6 and 4B3 are two monoclonal antibodies (mAb) which have originally been isolated as IgG1 isotype from seroconverted HIV-1 patients and bind to the principal immunodominant domain of gp41. In the course of this project both mAbs were isotype switched to IgA1. Recombinant CHO cell lines were established for the production of 3D6 and 4B3 as dimeric as well as secretory IgA. While dIgA were expressed by a single cell line, sIgA are produced by a biochemical association of dIgA with SC. Both dIgA and sIgA variants were characterized and the contribution of the heavily glycosylated SC on IgA stability will be investigated.

Background and aims Anti-idiotypic antibodies (Abs) represent an alternative vaccination approach in human therapy. This approach is based on the idiotype (Id) network theory postulated by Jerne describing the Ab (Ab1) – anti-idiotypic Ab (Ab2) – anti-anti-idiotypic Ab (Ab3) cascade stimulation. Specific anti-Id Abs serve as an “internal image” of the target antigen and can be used to induce Abs able to bind to the cognate antigen [1]. The anti-Id Ab Ab2/ 3H6 [2], developed at our Institute was generated in mouse and is directed against the human monoclonal Ab (mAb) 2F5, which broadly and potently neutralizes primary HIV-1 isolates [3]. Ab2/3H6 which has been characterized previously [4,5] is able to mimic the antigen recognition site of 2F5 and therefore it is suggested as a putative candidate for an HIV-1 vaccine. We investigated the potential of Ab2/3H6 by immunization of Fab fragments and fusion proteins with interleukin 15 (IL15) and tetanus toxin (TT) tags as immune modulators. After three prime/boost administrations rabbit sera were purified and analyzed for 2F5-like specific Abs. Further, the 2F5-like Abs from the sera were enriched by affinity purification and characterized for their binding affinity to 2F5. In an additional approach we applied different humanization methods to reduce the immunogenicity of the originally mouse derived Ab2/3H6. The mouse variable regions of Ab2/3H6 were subjected to three different humanization methods, namely resurfacing, CDR-grafting and superhumanization. Four differently humanized Ab2/3H6 variants were characterized for their binding affinity to 2F5 in comparison to the original Ab2/3H6. Results To evaluate the humoral immune response of Ab2/3H6 we designed Ab2/3H6 fusion proteins with IL15 and TT. Recombinant CHO cell lines were established and after protein purification New Zealand white rabbits were immunized with the Ab2/3H6 variants. Ten days after the final boost sera were collected and analyzed for total rabbit IgG levels. After proteinA affinity purification of the sera the isolated rabbit IgGs were tested for Ab2/3H6 and recombinant gp140 (UG37) specificity (Figure 1A). Further an affinity enrichment step using a UG37/ELDKWA column was performed and the obtained Ab3 fraction was tested on UG37 (Figure 1B) and additionally on the original 2F5 epitope ELDKWA (Figure 1C). Finally the Ab3 fraction was tested for binding affinity to the UG37 in a bio-layer interferometry assay which showed that the Ab3 fraction has a 6.6 fold reduced affinity towards UG37 compared to the mAb 2F5 (Table 1). For the humanization approach three different methods were chosen. The “resurfaced” variant (RS3H6) was developed by a computer model and surface exposed amino acids in the murine framework (FR) were substituted by residues usually found at equivalent positions in human Abs. The “superhumanized” form (SH3H6) was designed by structural homologies between the murine Ab2/3H6 CDRs and human germline CDRs. The most homologous human germline Ab was then used as acceptor FR. For the “CDR-grafted” variant two versions were expressed. An “aggressive” graft (GA3H6) harbouring less backmutations, making the grafted Ab more human-like and a “conservative” graft (GC3H6) with more backmutations. The different “reshaped” variable regions of Ab2/3H6 were expressed in CHO-cells as IgG1 molecules. The obtained “humanized” Ab variants were further characterized by competition ELISA (Figure 1D). * Correspondence: renate.kunert@boku.ac.at Department of Biotechnology, Institute for Applied Microbiology, BOKU – University of Natural Resources and Life Sciences, A-1190 Vienna, Austria Kunert and Mader BMC Proceedings 2011, 5(Suppl 8):P64 http://www.biomedcentral.com/1753-6561/5/S8/P64