Process development and cell engineering for enhancement of human bone morphogenetic protein-4 production in recombinant CHO cells

Abstract

rhBMP-4 has a potential for therapeutic uses for bone and cartilage repair, and treatment of osteoarthritis and rheumatoid arthritis through its ability to stimulate chondrogenesis. For clinical applications, obtaining large quantities of biologically active rhBMP-4 is essential. rhBMPs used in clinical applications are produced in Chinese hamster ovary (CHO) cells. rhBMP-4 is synthesized as large precursors, which then undergo proteolytic cleavage within the secretory pathway by proprotein convertases (PCs) to produce functionally active mature dimers. To improve product yield, several strategies such as PC overexpression and optimization of the cell culture environment have been attempted with limited success.

For large scale production of rhBMP-4, a recombinant CHO cell line stably expressing hBMP-4 was established by dhfr/methotrexate (MTX)-mediated gene amplification. The rhBMP-4 concentration reached a maximum at 4 µg/mL on day 3-5, and then decreased along the cell viability until the end of culture. Incubation of the supernatants obtained in mid-exponential phase (day 4) with cells obtained in death phase (day 6) indicated that a combination of the action of proteases and glycosidases released from the dead cells could be responsible for decrease of rhBMP-4 concentration at the end of cultures. To maximize the production of rhBMP-4 from CHO cells, the effects of culture temperature and pH on cell growth and BMP-4 production were investigated. Cells were cultivated in a 2L-bioreactor at different culture temperatures (33 and 37℃) and pH (6.7, 6.9, 7.2, and 7.5). Lowering the culture temperature arrested cell growth and increases the specific productivity (qrhBMP-4). The highest rhBMP-4 concentration (4.6 ± 0.5 μg/mL) was obtained at pH 7.2 and 37 °C. The extent of decrease of rhBMP-4 concentration was alleviated as maintaining high cell viability for a longer culture period at the lower culture temperature. Regardless of culture temperature and pH, significant rhBMP-4 degradation via increased proteolytic activity was observed in the decline growth phase. Protease inhibitors such as EDTA and leupeptin were toxic and thus did not enhance rhBMP-4 production. High viability was achieved in repeated batch cultures with frequent medium exchanges. These cultures showed enhanced rhBMP-4 productivity, demonstrating the potential of repeated batch cultures as a means for improved rhBMP-4 production.

Endocytic regulation serves a critical role in modulating the extracellular level of signaling molecules including BMPs. Unfortunately, endocytosis may result in poor yields of rhBMP-4 from CHO cell cultures. When rhBMP-4 was incubated with CHO cells, rhBMP-4 was actively internalized into cells. Cell surface bound-heparan sulfate proteoglycans (HSPGs) served as the major receptors for rhBMP-4 internalization. Removal of cell surface heparan sulfate (HS) by heparinases or reduction of HSPG synthesis by knockdown of xylosyltransferase2 (xylt2) in CHO cells decreased internalization of rhBMP-4. In addition, treatment with endocytosis inhibitors (chlorpromazine, genistein, and dynasore) identified a clathrin- and dynamin-dependent endocytic pathway as the major route for rhBMP-4 internalization. To enhance product yield by minimizing rhBMP-4 internalization in recombinant CHO (rCHO) cell cultures, we have tested various strategies to reduce HSPG synthesis (knockdown of xylt2 and sodium chlorate treatment) or inhibit the binding of rhBMP-4 to cell surface-bound HSPGs (supplementation with heparin or dextran sulfate (DS)). Among these approaches, DS, which is a linear anionic sulfated polysaccharide with similarity to HS chains, was the most effective in enhancing rhBMP-4 production in rCHO cell cultures. Compared with the control cultures, DS addition to culture medium (1.0 g/L) resulted in 1.4-fold and 2.3-fold increases in maximum rhBMP-4 concentration in batch and fed-batch cultures, respectively. Thus, rhBMP-4 in culture medium is internalized into CHO cells through cell surface-bound HSPGs and the addition of DS to the culture medium significantly increased rhBMP-4 production by blocking the internalization of rhBMP-4 into rCHO cells.

As CHO DG44 host cells express the essential elements of BMP signaling pathways, a CHO cell line expressing rhBMP-4 is capable of autocrine BMP signaling and hence having the unexpected cellular functions affected by rhBMP-4 mediated signaling pathways. The previous works in other cell lines suggest that the negatively impacted cell growth and the auto-downregulation of BMP-4 gene including the post-transcriptional regulation on BMP-4 mRNA and it could contribute to the low rhBMP-4 productivity in CHO cells. In this study, we observed that treatment of a selective inhibitor of the BMP type I receptors, LDN-193189, increased hBMP-4 mRNA expression and rhBMP-4 production. To clarify the effect of rhBMP-4, we investigated the inhibitory effects on cell culture profiles and the global changes of protein expression regarding cell growth and apoptosis upon rhBMP-4 treatment in CHO DG44 host cells. Using CRISPR/Cas9 gene-editing technologies, we knocked out BMPRII or BMPRIA gene in CHO DG44 cells to prevent the BMP-induced signaling. Using three different KO host cell lines (BMPRIA-KO-47, BMPRII-KO-20, BMPRII-KO-118) and a DG44 wild-type (wt) cell line, rCHO cell clones producing hBMP-4 were generated by a stepwise selection with increasing methotrexate (MTX) concentrations. The BMP-unresponsive BMPR KO host-derived cell lines resulted in increased hBMP-4 mRNA expression levels and improved productivity compared to the wild-type (wt) cell lines. The top high producing KO-derived and wt-derived clones were selected and evaluated with respect to the hBMP-4 mRNA levels and the rhBMP-4 productivities during the batch cultures. Using the knockout host cells, we generated hBMP-4 producing CHO cells free from autocrine BMP signaling and achieved the highest rhBMP-4 productivity (35.0 ± 0.9 μg/mL).