Experiences of virus, retrovirus and retrovirus-like particles in Chinese hamster ovary (CHO) and hybridoma cells used for production of protein therapeutics.
Garnick and coworkers indicated that they experienced two independent MVM outbreaks in a period where approximately 2000 fermentations were performed, hypothesizing that such events were rare but inevitable consequences of very large scale operations. In GIs experience over the last 12 years we have seen no incidence of MVM (or any other virus) in close to 3000 fermentations, albeit at lower volumes than produced at Genentech; GI has used 250-2500L bioreactors for manufacturing whereas Genentech have reported using 100-10,000L bioreactors. Nonetheless, volumes of complex media in the same range as used at Genentech have been used at GI with no observations of viral contamination events. The reason for this is not clear. However, GI's experience in combination with experience from sub-contract testing agencies who service the majority of the biotechnology industry may call the inevitability of an MVM outbreak into question. It would appear that very few adventitious viral contaminations of cell cultures have occurred in industry in the last decade. Interestingly, the frequency of contamination events appear to be lower in CHO cells than in hybridoma cells. It should be noted, however, that these conclusions are not statistically based and the scope of the above survey was somewhat limited. RVLPs are present in both CHO and hybridoma cells. The characteristics of both are compared in Table 4. C-type particles from hybridoma cells are more abundant as a rule than those from CHO cells. Although the majority of C-type particles produced by hybridoma cells appear to be non-infective (in S+L- assays), approximately one in a million particles are competent to replicate in S+L- cells. The evidence that C-type particles can replicate in human cells has proved difficult to reproduce consistently. It is likely that replication of xenotropic hybridoma C-type particles in human cells is inefficient or restricted to only a small number of specific cell lines. C-type RVLPs from CHO cells are produced less abundantly than those from hybridoma cells and are not competent to replicate due to a defective endonuclease gene. However, over the last two decades the use of hybridoma cells and products derived from these cells has not provided any evidence of transmission of these viruses to humans; in addition they can be readily removed or inactivated. Thus, neither agent would appear to constitute a significant risk to pharmaceutical products made from their respective host cells. Nonetheless, given the difference in relative safety profiles between RVLPs from CHO and hybridoma cells it is not unreasonable to propose that safety factors (clearance factors in removal/inactivation studies in excess of the reduction of virus loads to zero) required should be less for a CHO process than for a hybridoma process.