Introduction
Recombinant proteins for human therapeutic use has increased
significantly in the last 25 years. Disulfide bonds play an important
role in maintaining and stabilizing the three-dimensional structure of
those proteins (Dombkowski et al., 2014; Wang et al., 2015). Their
reduction was observed now and then during a recombinant protein
manufacturing (Trexler-Schmidt et al., 2010;
Kao et al., 2010; Mullan et al., 2011). Such a reduction event was
catastrophic as the product failed to meet the drug substance
specifications and the bulk product was lost. The reduction has a
substantial impact on the functional effects of a mAb, including
variable effects on antigen binding and Fc function, with the potential
to significantly impact mAb efficacy in vivo (Gurjar et al., 2019). The
reduction mainly took place in the harvested cell culture fluids (HCCF).
The release of intracellular reducing components was widely considered
to be the cause of the observed reductions
(Trexler-Schmidt et al., 2010; Kao et al.,
2010). The release took place both in cell culture bioreactors and in
the harvest process of cell culture fluids (CCF). However, the harvest
process was widely considered to be major
(Trexler-Schmidt et al.,2010; Kao et al.,
2010). At the end of a production phase, the cell culture fluid was
usually harvested either by disc stack centrifugation followed by depth
filtration or by depth filtration only. The mechanical stresses from
centrifugation and depth filtration were reported to cause cell lysis.
The disulfide bond reduction of monoclonal antibodies has been studied
intensively. The thioredoxin reduction pathway, including thioredoxin
reductase (TrxR) and NADPH, has been proposed as a primary contributor
(Kao et al., 2010; Koterba et al., 2012). The monoclonal
antibodies contain both intra chain disulfide bonds and inter chain
disulfide bonds. The inter chain disulfide bonds are more susceptible to
reduction than intra ones (Liu et al., 2012). For the two studied IgG1,
the disulfide bonds between the light chain and heavy chain were more
susceptible for reduction than those between two heavy chains. The upper
disulfide bond of the two inter heavy chain disulfide bonds was more
susceptible than the lower one (Liu et al., 2010). Reduction-susceptible
disulfide bonds are frequently located at the surface of a protein (Wang
et al., 2015). There were large differences in reduction sensitivities
between different monoclonal antibodies (Hutterer et al.,2013; Wang et
al., 2015). Moreover, the reduction power in CCF and HCCF could vary
largely for different molecules and different processes as well
(Hutterer et al., 2013). The reduction occurrence depends on molecule,
cell culture process, and harvest process.
The various measures to prevent the disulfide reduction occurrence had
been developed. Those measures included air sparging, air overlay, pH
adjustment, cooling down, TrxR inhibitors and chemical additions. The
added chemicals could be H2O2, EDTA,
Cu2+,Hg2+, Zn2+,
Co2+, Mn2+, and so on
(Trexler-Schmidt et al.,2010; Chung et al.,
2017; Mun et al., 2015; Chaderjian et al., 2005; Du et al, 2018). Among
them, air sparging or air overlay is most often used due to its
simplicity and minimal potential impact on product qualities (Mun et
al., 2015). The implementation of the above measures in a recombinant
protein manufacturing will inevitably increase its operation complexity
somewhat and might impact its product qualities as well sometimes. A
decrease in main peak and an increase in acidic peak were reported when
air sparging was applied during a HCCF hold for a given monoclonal
antibody (Mun et al., 2015). Such a decrease in main peak and an
increase in acidic peak had been seen in our labs as well for some
molecules (data not shown here). Besides that, the decrease in the
filterability of HCCF was observed in our labs and at-scale
manufacturing (data not shown here) when air sparging or air overlay was
applied.
If air sparging only doesn’t work, the cooling HCCF down plus air
sparging might be worth to be considered next. In the case that the
cooling down plus air sparging was still not sufficient, the addition of
a weak oxidant, such as Cu2+, might be a choice.
Certainly such an addition would introduce additional complexity to a
process. It is highly desirable to apply a reduction measure only if
necessary and to apply no strong measure if a moderate one works. How to
assess the level of the risk of disulfide bond reduction for a given
process quantitatively? How to determine what measure is appropriate for
a given process? This study reported a quantitative method to assess the
risk of disulfide bond reduction for a given process. The outcome of the
assessment could tell whether a reduction measure should be taken and
what measure is appropriate if yes. Answer what measure (mild, moderate,
or strong) shall be appropriate for a given process.