A new Chinese hamster ovary cell line expressing alpha2,6-sialyltransferase used as universal host for the production of human-like sialylated recombinant glycoproteins.

Chinese hamster ovary (CHO) cells are widely employed to produce glycosylated recombinant proteins. Our group as well as others have demonstrated that the sialylation defect of CHO cells can be corrected by transfecting the alpha2,6-sialyltransferase (alpha2,6-ST) cDNA. Glycoproteins produced by such CHO cells display both alpha2,6- and alpha2,3-linked terminal sialic acid residues, similar to human glycoproteins. Here, we have established a CHO cell line stably expressing alpha2,6-ST, providing a universal host for further transfections of human genes. Several relevant parameters of the universal host cell line were studied, demonstrating that the alpha2,6-ST transgene was stably integrated into the CHO cell genome, that transgene expression was stable in the absence of selective pressure, that the recombinant sialyltransferase was correctly localized in the Golgi and, finally, that the bioreactor growth parameters of the universal host were comparable to those of the parental cell line. A second step consisted in the stable transfection into the universal host of cDNAs for human glycoproteins of therapeutic interest, i.e. interferon-gamma and the tissue inhibitor of metalloproteinases-1. Interferon-gamma purified from the universal host carried 40.4% alpha2,6- and 59.6% alpha2,3-sialic acid residues and showed improved pharmacokinetics in clearance studies when compared to interferon-gamma produced by normal CHO cells.

Recombinant HMG1 protein produced in Pichia pastoris: a nonviral gene delivery agent.

This paper describes the production of a recombinant protein from the expression system based on the methylotrophic yeast Pichia pastoris. Efficient production of rat high-mobility-group 1 (HMG1) protein was obtained using the system. Two forms of HMG1 were secreted into the culture medium: a 24.5-kDa species corresponding to the native HMG1 and a 32-kDa glycosylated derivative. Non-glycosylated recombinant HMG1 was purified easily and shown to possess the same DNA-binding properties as HMG1 purified from calf thymus. Plasmid DNA complexed to the recombinant HMG1 is taken up by a variety of mammalian cells in culture. Transient expression of a luciferase reporter gene was observed. Under selective conditions, stable expression of a neomycin gene was established as a result of integration into the genome. HMG1-mediated gene delivery was as efficient as calcium phosphate-mediated transfection but without associated cell damage. In addition, stable transfectants obtained after selection for G418 resistance usually integrated only one copy of the transfected DNA in contrast to the high unpredictable number obtained by the calcium phosphate method. HMG1 transfection complexes were not toxic to cultured cells, even at high concentrations.

Genetic engineering of α2,6-sialyltransferase in recombinant CHO cells and its effects on the sialylation of recombinant interferon-γ.

The CHO cell line has achieved considerable commercial importance as a vehicle for the production of human therapeutic proteins, but is known to lack a functional copy of the gene coding for α2,6-sialyltransferase (EC 2.4.99.1). The cDNA for rat α2,6-ST was expressed in a recombinant CHO cell line making interferon-γ, using a novel in vitro amplification vector. The enzyme was expressed efficiently, and resulted in up to 60% of the total sialic acids on interferon-γ being linked in the α2,6-conformation. This sialic acid linkage distribution was more akin to that seen in natural human glycoproteins. In the most successful cell clones, expression of α2,6-sialyltransferase improved the overall level of sialylation by up to 56%, and had no adverse effects on cell growth, IFN-γ productivity or other aspects of IFN-γ glycosylation. These experiments demonstrate how the glycosylation machinery of rodent cells can be genetically manipulated to replicate human tissues.

Characterisation of biotinylated liposomes for in vivo targeting applications.

Liposomes containing monosialoganglioside (GM1) or polyethylene glycol (PEG) lipid derivatives have prolonged circulation in the blood. This favours liposome extravasation to tumour sites. In this report it is shown that inclusion of GM1, PEG550-DPPE or PEG2000-DPPE in liposomes containing biotin-DPPE significantly diminished the ability of vesicles to bind to streptavidin in vitro. Steric inhibition due to the bulky head group of these lipids was least for biotin-DPPE liposomes containing GM1. Biodistribution studies in C26 tumour-bearing mice showed that GM1-liposomes containing small amounts of biotin-DPPE have long circulation life-times in the blood. Using fluorescent microscopic techniques, liposomes containing both GM1 and biotin-DPPE were detected within extra-vascular spaces in tumours. In addition it was shown that biotin-DPPE in GM1-liposomes bound streptavidin in situ. These results suggest that GM1-liposomes containing biotin-DPPE have potential use as diagnostic or therapeutic reagents in pre-targeting applications dependent on the high-affinity interaction of biotin with streptavidin.

Postsurgical adjuvant chemoimmunotherapy with recombinant interleukin-2 and 1,3-bis-(2-chloroethyl)-1-nitrosourea on spontaneous metastases of a non-immunogenic murine tumour.

The efficacy of the association of recombinant interleukin-2 (rIL-2) with chemotherapy has been investigated on an experimental model representative of clinical tumours, i.e. on post-surgical spontaneous metastases of a non-immunogenic tumour. We used the M5076 ovarian reticulum cell sarcoma, which metastasizes to the liver after intra-footpad implantation. Such a tumour appeared to be non-immunogenic by a variety of commonly used in vivo assays. Four clinically widely employed drugs, i.e. doxorubicin, cis-diamminedichloroplatinum II, cyclophosphamide and 1,3-bis-(2-chloroethyl)-1-nitrosourea (BCNU), were tested and BCNU proved to be the most effective one when administered as single injection at the maximum tolerated dose (33 mg/kg i.p.) 1 day after tumour excision. When moderate doses of rIL-2 (6 x 10(5) IU in three injections per day for 5 days) were administered at three different intervals after BCNU, namely before the nadir of white blood cells (1 day after BCNU), at the nadir (3 days after BCNU) or at recovery (6 days after BCNU), no increase in BCNU antitumour activity was observed. The same results were obtained by administering rIL-2 for 5 days before BCNU. Higher doses of rIL-2 (1.2 x 10(6) IU in three injections per day for 5 days), which were always well tolerated in sham-excised non-tumour-bearing mice, proved lethal in two out of four experiments in tumour-bearing animals. In the two experiments in which no lethality was observed, the administration of high doses of rIL-2 1 or 6 days after BCNU significantly increased the antitumour activity of BCNU alone. rIL-2 alone was not active even when administered at high doses. These results indicate that high but not moderate doses of rIL-2 may increase the activity of BCNU against a non-immunogenic tumour. Moreover, they suggest that rIL-2 tolerability is reduced in tumour-bearing mice.